THE UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF THE COLLEGE OF ENGINEERING LIQUID AND GAt-PHASE MASS TRANSFER ON BUBBLE-CAP TRAYS John W. Begley A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the University of Michigan 1959 October, 1959 IP-395

An,~~~~I, Asters d't

ACKNOWLEDGMENTS The author wishes to gratefully acknowledge the assistance of the following persons during the course of this study: Members of the Doctoral Committee Professor Brymer Williams, Chairman Associate Professor T. C. Adamson Professor J. T. Banchero Assistant Professor K. -F. Gordon Professor R. R. White Professor J. L. York A.I.Ch.E, Plate Efficiency Research Assistants James R. Beissel Talivaldis Cepuritis Jon M. Gaston Kathryn Hilbert Richard C.Keezer Charles D. Malloch Constantinos T. INikoletopoulos Robert Norman Eugene D. Salesin Robert J. Van Duyne James Wall Chi Hua "Ruth" Wu Shop and Office Personnel in the Department of Chemical and Metallurgical Engineering Cleatis Bolen David M. Brown, Mass Spectrometer Lab. Frank Drogosz Ludwig Eppler George Foster William Hines Madelaine R. Ingerson Financial support for the research program was supplied by the A.I.Ch.E. Plate Efficiency Research Committee with funds solicited from approximately 35 industrial organizations. Fellowship grants: from Phillips Petroleum Company and the American Institute of Chemical Engineers were a direct assistance to the author during the studies and research work for the doctorate. The chemicals used in the studies were ii

supplied gratis by the Dow Chemical Company through the arrangements of Messrs. N. Poffenberger and P. C. Dean. Miss Rluth Wu, one of the research assistants, deserves special mention because of her interest in the research problems and her devotion to the tedious job of transposing the laboratory data to the form used herein. The author would also like to acknowledge the patient work of his wife, Lois Begley, during the preparation of the original draft for the thesis. iii

TABLE OF CONTENTS Page ACKNOWIEDGE1ENTS....................,.......... ii LIST OF TABLES 4P. **.0 f- w! Vi LIST OF TABLES - APPENDIX. a......,....*.......... vii LIST OF FIGURES.......*........,....,........... ix LIST OF FIGURES - APPENDIX....................... x NOMENCLATURE..................... xvi INTRODUCTION.............+ *. *. * * *.......... a.... 1 TRANSFER OF MATERIAL BETWEEN PHASES.. 0............,.. 9 BASIC RELATIONSHIPS BETWEEN POINT EFFICIENCY AND MASS TRANSFER COEFFICIENTS........ *............. 13 MURPHREE LIQUID POINT EFFICIENCY.... 20 BASIC RELATIONSHIPS BETWEEN POINT EFFICIENCY AND PLATE EFFICIENCY................. *,'''.......9.. 25 PREVIOUS INVESTIGATIONS OF THE EFFECT OF LIQUID PROPERTIES - VISCOSITY AND DENSITY *....................... 37 PURPOSE OF PRESENT INVESTIGATIO N....................* 45 SUMMARY OF EXPERIMENTAL INVESTIGATIONS.*...................... 48 MATERIALS...........*..............*...., 50 EQUIPMENT AND LABORATORY PROCEDURES......,.................. 52 Equipment.........5....,,.. 52 Laboratory Procedures - Vaporization Studies........ 57 Laboratory Procedures - Absorption Studies..........57 Laboratory Procedures - Hydraulic Data..... 58 PRESENTATION OF DATA................................ 62 Hydraulic Data - General Observations............... 62 Hydraulic Data - Gas and Liquid Holdup........... 66 Hydraulic Data - Relative Froth Density............. 81 Mass Transfer Data - Vaporization Studies......... * 93 Mass Transfer Data - Absorption Studies............. 101 iv

TABLE OF CONTENTS (CONT'D) Pae_ CORRELATION OF DATA. *,,......*,. O 114 Vaporization Studies..,,..... 114 Absorption Studies....... * 140 Mixing Studies...6..... *0 00 *,, *, * 161 DISCUSSION OF RESULTS.. 9 * * * * *. 9 *. * * 177 Vaporization Studi es.......... 177 Absorption Studies............,.,.............. 9*L Absorption Studies.l.,.. 191 CONCLUSIONS... * * *. 9 * * G i.a. + 9 9 o 9 *o0 * *O * 201 Hydraulic Characteristics...,............. 201 Gas-Phase Mass Transfer.., 202 Liquid-Phase Mass Transfer....9..P.,... 0*.*990#9* 203 Comparison Between Gas- and Liquid-Phase Mass Transf'er...... e. e..... 205 Liquid Mixing...*......,,...., * o.-................ 205 RECOMMENDATIONS,... 9 *0 e. * * 6* *o** * *,. ** o e t *9oe *o9e 206 Gas-Phase Mass Transfer,................... *...O O. 206 Liquid-Phase Mass Transfer......,..... 207 APPENDIX A - DESCRIPTION OF APPARATUS....................,... 208 APPENDIIX B - LABORATORY TECHNIQUES......................,... 226 APPENDIX C - ESTIMATE OF EXPERIMENTAL ERROR..#...o..**.......#*Q 240 APPENDIX D- CALIBRATION DATA.,....... 249 APPENDIX E - PHYSICAL PROPERTIES...,,.,,.,..,..,,,., 253 APPENDIX F - SAMPLE CALCULATIONS, t.,....,,..,,..*....9.260 APPENDIX G - EXPERIMENTAL AND CALCIULATED DATA,..,..... **.... 273

LIST OF TABLES Table Page I Definitions of N0G and NOL.......................... 24 II Relationships Between Point and Plate Efficiency...T 26 III Summary of Previous Efficiency Studies............. 39 IV Physical Properties of Dow Cyclohexanol............ 50 V Physical, Properties of Dow Ethylene Dibromide....*. * 51 VI Characteristics of the Tray Layout,................ 54 VII Ratio of Point Concentrations to Equilibrium Concentrations on Test Tray..*..,.... 111 VIII Summary of Graphical Analysis of Vaporization Data.. 117 IX Summary of Vaporization Data Correlation by Use of Equation (108) and Regression Analysis......... 118 X Partial Correlation Coefficients for the Dimensionless Groups in Equation (108)................. 122 XI Comparison of Diffusivities for Oxygen in WaterSucrose Solutions Predicted by Wilke-Chang Correlation and Experimental Data,............ 147 XII Solvent Abnormality Factors and Parachors............ 154 XIII Correlation of Liquid-Phase Mass Transfer Coefficients............ 156 XIV Fairbrother's Data for Desorption of Carbon Dioxide From Water and Water-Glycerol Mixtures.... 159 XV Sensitivity of Equation (67) to Changes in the Parameter t"C" at Various Values of XEo........... 164 XVT Correlation of Mixing Parameter "C".....,...... 170 ~~~~~~~~~ e~e ~ee e ~I~...

LIST OF TABLES - APPENDIX Table Page I-B Comparison of Sampling Methods.................... 236 I-C Percentage Error in EG and NG Resulting From 1% Errors in y*, Yl, and y....0..............0..... 243 II-C Relative Magnitude of the Terms in Equation (3-c).... 244 III-C Variation of the Percentage Error in Plate Efficienty, EML...0. o................o*e...... 248 I-D Calibration Data for Rotameter No. D6-2445, Calibration by Using Water.................... 249 II-D Calibration Data for Rotameter No. D6.-2445, Calibration by Rotary- Meter.............o*.....o..... * 250 III-D Calibration Data for Rotameter W70-4024/1,,Calibration by Using Wat:er.......................... 252 IV-D Calibration Data for Wet-Test Meters, H9SS and JSSo o oo o oeooo o o * o..... e.o. oe.. o.eeoe 252 I-E Viscosity of Cyclohexanol....................... 253 II-E Density of Cycelohexanol.......................,. -. 255 III-E Solubility Data for Carbon Dioxide in Cyc lohexanol.. o -q of.O.0a o9.. a 004 4 @N~. e.... a -.:.. - 257 IV-E Vapor Pressure of Cyclohexanoli...... ~....~...a...257 V-E Vapor Pressure of Ethylene Dibromide.............e.. 257 I-F Laboratory Data Sheet - Vaporization Data........... 265 II-F Laboratory Data Sheet - Absorption Data.............. 266 I-G Plate Efficiencies in Rectangular Column at University of Michigan (Original Data for N2Ethylene Dibromide System) 1.50 in. Weir, 2.0 in. Splash Baffle....,.................... 274 II-G Prediction of Vaporization Data.............,. 291 vii

LIST OF TABLES - APPENDIX (CONT' D) Table Page III-G Plate Efficiencies in RectangUlar Column at University of Michigan, Carbon DioxideCyclohexanol System Weir Height, 3-1/2 in. Splash Baffle Height, 4 in.,,..,..-.....,........ 297 IV-G Mass Transfer Units and Mass Transfer Coefficients - Absorption Studies.,. @................ 304 V-G Warzel's Data for Carbon Dioxide Absorption and Iesorption Using Air-Water System With Gas Phase Dilute in Carbon Dio xide...................... 307 VI-G Comparison of Point Efficiency Values Determined By Use of Several Different Relationships Between Point and Plate Efficiency................. 309 VII-G Data Used in the Correlation of the Avera'ge Liquid Concentration on the Tray...............,., 312

LIST OF FIGURES Figure Ige 1 Mixing Model Used to Derive Relationships Between Point and Plate Efficiency....................... 27 2 Front View of Test Column and Auxiliary Equipment.... 53 3 Single Tray in Absorber, Showing Removable Deck and Adjustable Splash Baffle.............55 4 Location of Sampling Points on Test Tray - Points, 2, 3, 4, and 5 Used Also to Measure Hydrostatic Head at Respective Points..........,..,.... 59 5 Top View of Removable Tray, Showing Bubble Caps and Location of Liquid Sampling Points............. 60 6 Bottom View of Removable Tray, Showing Inlet to Risers and Liquid Sumps Below Sampling Point....... 60 7 Froth Holdup on Test Tray, Carbon-DioxideCyclohexanol System; Weir Height, 3-1/2 in.; Splash Baffle Height, 4 in.; F-Factor, 0.62; Liquid Viscosity, Approximately 55 cp...*.......... 63 8 Froth Holdup on Test Tray, Carbon-DioxideCyclohexanol System; Weir Height, 3-1/2 in.; Splash Baffle Height, 4 in.; F-Factor, 1.1; Liquid Viscosity, Approximately 55 cp...*....*... 64 9 Clear-Liquid- Height Data for Nitrogen-Cyclohexanol System. e *................... * * * 68 10 Clear-Liquid-Height Data for Nitrogen-Ethylene Dibromide System.,,,.................,.............,.69 11 Clear-Liquid-Height Data for Nitrogen-Cyclohexanol System at High Gas Rates........... 70 12 Froth-Height Data for Nitrogen-Cyclohexanol System. Data are Presented as a Function of Superficial Gas Velocity with Parameters of Weir Height, Liquid Viscosity and Liquid Rate....... 72 13 Froth-Height Data for Nitrogen-Ethylene Dibromide System. Data are Presented as a Function Superficial Gas Velocity with Parameters of Weir Height and Liquid Rate,........................ 73 ix

LIST OF FIGURES ( CONT D ) Figure Page 14 Froth- eight Versus F-Factor-Carbon DioxideCyclohexanol System. Variable Liquid Viscosity 2 in, Weir; 2-1/2,in. Splash Baffle..,..,.,...., 74 15 Froth-Height Data for Several Systems Used in Vaporization Studies. Data are Presented as a Function of Superficial Gas Veloeity with Parameters of Liquid Viscosity anad Density and. Gas Density; 1-1/2 in. Weir Height; 8.0 GPMo,......... 76 3. 76 16 Froth-Height Data for Several Systems Used in Vaporization Studies, Data are Presented as a Funetion of F-Factor with Parameters of Liquid Viscosity and Density; 1-1/2 in. Weir Height; 8 0 GPM[. a 6 o a a 0 o* a oo o o o * 0 o b O * a 77 17 Gas Hotld-Up Data for Nitrogen Cyclohexanol System, Data are Presented as a - Function of uperficial Gas Velocity with Parameters of Weir Height, Liquid Viscosity, and Liquid Rate.,,,.. g......,., 78 18 Gas Hold-Up Data for Nitrogen-Ethylene Dibromide System -Data Presented as a Function of Superficial Gas Velocity with Parameters of Weir Height and. Liquid Rate,.O,.. f 0 0 O*.0.0Q. 79 19 Gas Hold-Up -Data for Several Systems. Data are Presented as a Funetion of F-Factor with Parameters of LiEquid Viscosity and -Density and Weir Height; 8.0 GPM,.,,,.,.,,. o o,,,,,, o,,O 80 20 Gas Hold-Up Data for Carbon Dioxide-Cyclohexanol and Air-Water Systems as a Function of F-Fator... 82 21 Gas Hold-Up Data for Carbon Dioxide-Cyclohexanol and Air-Water Systems as a Function of F-Factor, O.. 83 22 Gas Hold-Up Data for Carbon Dioxide-Cyclohexanol and Air-Water Systems as a Function of F-Factor,,o, 84 23 Gas Hold-Up Data for Carbon Dioxide-Cyclohexanol and Air-Water Systems as a Function of F-Faetor.,o 85 24 Relative Froth Density for Several Systems. Data Presented as a Function of F-Factor with Parameters of Weir Heig ht: Liquid. Rate, and Liquid Prope;rties D ff, o. Qeo oo *o oo*OO 87, x

LIST OF FIGURES (CoNT' D) Figure Page 25 Relative Froth Density for Several Systems. Data are for 1-1/2 in. Weir and Constant Liquid Rate of About 8.0 GPM..................... 88 26 Relative Froth Density for Several Systems. Data are for 2 in. Weir, Variable Liquid Rate and Liquid Properties,..................................89 27 Relative Froth Density for Several Systems. *Data are for 3-1/2 in. Weir, Variable Liquid Rate and Liquid Properties............ *...... 90 28 Comparison of Relative Froth Density on the Tray Used in the Present Investigation and the Density.Reported by Gerster for a Similar Tray with No Splash Baffle............ 92 29 Gas -Phase Efficiencies for Nitrogen —Cyclohexanol System. -Data are Pres.ented as a Function of Superficial Gas Velocity with Parameters of Weir Height, Liquid Rate, and Liquid Viscosity...... 94 30 Gas-Phase -Efficiencies for Nitrogen-Ethylene Dibromide System. Data are Presented as a Function of Superficial Gas Velocity with Parameters of Weir Height and Liquid Rate......... 96 31 Number of Mass Transfer Units for Nitrogen - Cyclohexanol System. Data are Presented as a Function of Superficial Gas Velocity with Parameters of Weir Height, Liquid Rate, and Liquid Vis osity......... o e * o 99 32 Number of Mass Transfer Units for Nitrogen - Ethylene Dibromide System. Data are Presented as a Function of Superficial Gas Velocity with Parameters of Weir Height and Liquid Rate......... 100 33 Murphree Liquid Efficiencies for Carbon Dioxide - Cyclohexanol System - Variable Liquid Rate and Liquid Viscosity - 3-1/2 in. Weir, 4 in. Splash BaSff.le.*o*.. 103 34 Murphree Liquid Efficiencies for Carbon Dioxide - Cyclohexanol System - Variable Liquid Rate and Liquid Viscosity 2 in, Weir, 2-1/2 in. Splash Baffles....*.*.**.4a o 104 xi

LIST OF FIGBE S (C ONTt D) Figure Page 35 Liquid Concentrations on Tray Floor - Carbon Dioxide - Cyclohexanol System.......... a.. a.. 107 36 Number of Liqcuid-Phase Mass Transfer Units for Carbon Dioxide - Cyclohexanol System - Variable Liquid Rate and Liquid Viscosity - 3-1/2 in. Weir, 4 in, Splash Baffle....O....,,',...... 109 37 Number of Liquid-Phase Mass Transfer Units for Carbon Dioxide - Cyclohexanol System - Variable Liquid Rate and Liquid Viscosity - 2 in -Weir, 2-1/2 in. Splash BaffleOOO,,oOOO O..,,.,....c.... 110 38 Ratio of Point Concentrations and Equilibrium Concentration on Test Tray; 3-1/2 in, Weir Height; -Liquid Rate, 4*95 GPM.......o.... ~. 0 112 39 Mass Transfer Coefficients for Nitrogen-Cyelohexanol System at Three Different Weir Heights - Liquid Viscosity = 12 O acp...... f o I.. 4 8. 126 40 Mass Transfer Coefficients for Nitrogen-Ethylene Dibromide System at 1-1/2 and 3-1/2 in. Weir Height....0........ o o o..........O. o...*..... * e. 1 27 41 Number of Mass Transfer Units for Ammonia-Water System Data of Warzel,.......o...... a........129 42 The Effect of Liquid Properties Upon the Relationship Between NG/(Zf- Zc)0'2 and F-Factor,........ 130 43 Correlation of the2Constant s'a" in the Equation NG/(Zf Zc)0' = C'" pGa/2usa with Kinematic Liquid Viscosity,.....,....... a.......... 132 44 Correlation of "n't in th- Equation KGa - GC" y a/(zf - zc)O2 pG1/2 with Liquid Absolute -Viscosity and Liquid Kinematic Vi$soSity....4...0................................. 133 45 Correlation of "C"t in the Equation ka = C>'pGn/2usl+a=n/(Zf - Zc)0.28pg1/2 with Physical Properties of Gas and Lqu aid...... o a.....o o 137 xii

LIST OF FIGURES (CONT'D) Figure Page 46 Correlation of the Constant "C" in the Equation kGa' C usl+a=n/(Zf - Zc)028 with Physical Properties of Gas and Liquid...................139 47 Mass Transfer Coefficients for Carbon DioxideCyclohexanol System. Variables: Gas Rate, Liquid Rate, Weir Height, and Liquid Viscosity............ 142 48 Correlation of Mass Transfer Coefficients for Carbon-Dioxide-Cyclohexanol System................. 143 49 Correlation of Maps Transfer Coefficients for Carbon Dioxide-Water System................... 145 50 Comparison of Experimental Diffusivities and Diffusivities Predicted by Wilke-Chang Correlation and Arnold Equation...................... 150 51 Correlation of the Solvent Abnormality Factor in Arnold's Equation with the Parachor for Several Organic Compounds, Water, and Aqueous Solutions.... 152 52 Liquid Diffusivities for Carbon Dioxide-Cyclohexanol System - Comparison of Values Predicted by use of Arnold's Equation with Experimental Values of Schoenborn. o.................. 153 53 Correlation of the Ratios of the Liquid Mass Transfer Coefficients and Diffusivities at a Constant F-Factor with Liquid Kinematic Viscosity - Data for Water and Cyclohexanol Systems and Water-Glycerol Solutions.......................... 158 54 Comparison of Point Efficiencies Predicted by use of "C" Correlation and Warzel's Mixing Equation with Experimental Point Efficiencies,...6......... 167 55 Comparison of Point Efficiencies Predicted by use of Warzel's "C" Correlation and Point Efficiencies Predicted use of New "C" Correlation........... 168 56 Comparison "C" Values Used by Warzel and Values Predicted by "C"t Correlation C02 - H20 System...... 169 57 Correlation of the Average Liquid Concentration on the Tray with XEOG. q. o,................. 172 xiii

LIST OF FIGURES (CONT' D) Figure Page 58 Variation of Xavg - X/(XEoG)O12 with F-Factor...... 173 xo- x 59 Comparison of Experimental Point Efficiencies with Point'Efficiencies for Complete Mixing and Plug Flow.. 0 0.. 175 xiv

LIST OF FIGURES - APPENDIX Figure Page 1-A Bubble Cap Plate Layout.......................... 209 2-A Removable Trays, Showing Bubble-Caps and Risers for Absorber...................................... 210 3,A Position of Probe for Outlet Vapor Sample........,,,.... 212 4-A Apparatus Used to Obtain Liquid Samples and Measure Hydrostatichead at Four Points on Tray Floor...................*...... 217 5-A Flow Diagram for Vaporization Studies............ 220 6-A Flow Diagram for Absorption Studies............... 221 1-B Humidity Chart for the Nitrogen-Cyclohexanol System Prepared by C.H, Wu........... 228 2-B Humidity Chart for the Nitrogen-Ethylene Dibromide System Prepared by C.H. Wu. 229 1-D Calibration of Rotameter D6-2445...........,. 251 1-E Viscosity of Cyclohexanol.................. 254 2-E Density of Cyclohexanol............... 256 3-E Henry's Law Constants for Carbon DioxideCyclohexanol System........................... - 258 4-E Henry's Law Constants for Carbon DioxideCyclohexanol System............... 259 xv

NOMENCIATURE a Constant in correlation equations a Interfacial area, ft2/ft3 of gas holdup a Interfacial area, ft2/ft3 of liquid holdup aI Interfacial area in the bubble formation zone, ft2/ft3 of gas holdup a2 Interfacial area in the froth and entrainment zones, ft2/ft3 of gas holdup a' Interfacial area per plate per unit time of gas flow a' Interfacial area, ft2/(sq. in. slot area)(inches of liquid depth) A Constant in correlation equations AA The abnormality factor for the solute in Arnold's equation AB The abnormality factor for the solvent in4 Arnold's equation A51 Total slot area per plate, sq. in. AT Active cross-sectional area of the tray, ft2 (0,615 ft2 for tray used in present study). b Constant in correlation equations B Tray width, ft. B Constant in correlation equations BF Thickness of liquid film in wetted-wall tower, ft. c Constant in correlation equations C Concentration in liquid phase, lb moles/ft3 C Constant in correlation equations "C" Warzel mixing parameter C* Equilibrium liquid concentration, Ib moles/ft3 C'' Constant in correlation equation xvi

NCLACATURE (CONT D) C1, C2 Constants in the solution to the differential equation which describes the concentration in the liquid as a function of distance along the tray. C02 Molar density of liquid-phase, lb moles/ft3 C2m fThe average concestration of the solvent in the liquid film, lb moles/ft. Ci Concentration in liquid phase at the interface, lb mole/ft3 Cp Specific heat of gas, BTU/1b mole-F p CR Rotameter calibration factor C2 C3 Concentration in liquid phase at four different points on C4, C5 the tray floor, lb moles/ft3 d Differential operator d Constant in correlation equations db Bubble diameter, ft. do Orifice diameter, ft. D tube diameter, ft. 2 D3E Eddy diffusivity, ft /hr. DG Gas diffusivity, ft2/hr. DL Liquid diffusivity, ft2/hr.DLo Liquid diffusivity at 0~C ft2/hr DS Slot width, ft. e Constant incorrelation equations EOG Overall column efficiency EG Single-phase vapor point efficiency EL Single-phase liquid point efficiency EML Murphree liquid efficiency EMV Murphree vapor efficiency G~~~~~~~~~~~x~;

NOMENCLATURE (CONT'D) EOL Over-all liquid point efficiency f Constant in correlation equations f Friction factor F Energy loss due to friction, ft-lb force/lb mass. F-factor Usp/G, ft/sec l/ g Gravitational acceleration, ft/sec2. G Gas or vapor flow rate, lb moles/hr. Gs Superficial mass velocity, lb/ft2-hr. Gavg Average gas or vapor rates between a point above and below the tray, lb moles/hr, h Effective liquid depth, inches hG Gas-phase heat transfer coefficient, BTU/hr-ft2- ~F hL Slot submergence, ft. H Henryt's Law constant, p/c, atm-ft3/lb moles or p/x, atm/mol fraction H Holdup of gas in froth, ft3/ft3 HL Height of a liquid-phase mass transfer unit, ft. H.T.U. Height of a theoretical mass transfer unit JD J factor for diffusion kG Gas-phase mass transfer coefficient, lb moles/hr-ft -atm kL Liquid-phase mass transfer coefficient, lb moles/(hr-ft2) (lb moles/ft3) k'G Gas-phase mass transfer coefficient, lb moles/(sec-ft2) ( lb moles/ft3 ) kq Gas thermal conductivity, BTU/hr-ft2-.~F/ft kG'a Gas-phase mass transfer coefficient, sec-1 kLa liquid-phase mass transfer coefficient, sec-1 k'Glal gas-phase mass transfer coefficient in the bubble formation zone kG2a2 gas-phase mass transfer coef. in froth and entrainment zones xviii

1NOMENCIATURE (CONT'D) K tConstant KOG Over-all gas-phase mass transfer coefficient, lb moles/hr -ft2 atm. KOL Over-all liquidphase mass transfer -coefficient, lb moles/(hr-ft2 )(lb moles/ft5) Ko0a Over-all gas-phase mass transfer coefficient sec1 2L AVertical distance from top of bubble cap slots to top of Weir, ft. lo Crest of liquid over top of weir, fto le Height of slot, ft.o L liquid flow rate, lb moles/hr m Slope of the vaporwlia uid eauilibrium curve of y versus xo m Constant in. a correlation equation M Mixirng parameter LS/DEZcBPML M Constant in a correlation equation MA Solute molecular weight MB Solvent molecular Weight MG Molecular weight of gas n -Number of ideal mixing stages on a tray n Constant in a correlation equation nb Bubble frequency NA Rate of mass transfer, lb moles/hrf t2 NG Number of gas-phase mass transfer units NL Number of liquid-phase mass transfer units NS Number of bubble sources NOG Number- of overall gas-phase mass transfer units xix

NOMENCLATURE (CONT'D) NOL Number of over-all liquid-phase mass transfer units p* Equilibrium partial pressure in gas-phase, atm. 1Pi ~Partial pressure in gas phase at the interface, atm. PG Partial pressure in gas phase, atm, PBM Logrithmic-mean partial pressure of nondiffusing component, atm. P Total pressure, atm. tPT Total pressure drop across the tray, lb/ft2 qi Rate of circulation, ft3/hr QG Volumetric air flow rate per orifice, ft3/hr. QGT Total vQlumetric gas flow rate to the tray, ft3/sec. QLT Total volumetric liquid flow rate to the tray, ft /sec. s Rate of surface renewal st Distance along the tray, ft. S ZLength of tray, ft. t Liquid interfacial contact time, hr. tb Bubble contact time,- hr. tG Gas contact time, sec. tL Liquid contact time, sec. tGl Gas contact time in the bubble formation zone, sec. tG2 Gas contact time in the froth and entrainment zones T Temperature, ~F or ~K Wet bulb temperature, ~F Tas Adiabatic saturation temperature, ~FO us Superficial gas velocity, ft/sec. uavg Average superficial gas velocity between a point above and below the tray, ft/sec. xx

NOMEMCLATURE (CONT D) V Specific volume, ft3/1bo Vb Bubble velocity, ft/hro Vf Froth velocity, ft/seco VA Molal volume of the solute at normal boiling point, cc/gm-mole VB Molal volume of the solvent at normal boiling point, c/gm-mole W Fraction of distance across the tray W Work term in energy balance x fLiquid concaentration, mol fraction Equilibrium liquid concentration, mol fraction x' ~Liquid concentration at a point on the tray, mol fraction x Liquid concentration at the tray inlet, mol fraction Xe Liquid concentration at a point on the tray between the inlet downcomer and the first row of bubble caps, mol fraction Xi Liquid concentration at the gas!liquid interface, mol fraction xo Liquid concentration at the tray outlet, mol fraction xag Arithmetic average liquid oncentration o the tray, mole fraction XL Liquid filma thickness, ftd XB Solvent association parameter y Gas concentration, mol fraction Sy* Equilibrium gas concentration, mol:fraction y' Gas concentratian at a point on the tray, mol fraction Yi Gas c:acenLtration at the gas-liquid interface, mol fraction Yavg Averag gas concentration above the tray, mol fraction YZ Gas cncentration below t;he tray, mcl fraction

NOM CLATURE (CONT'D) Z Potential energy term, ft-lb force/lb mass Zc Average clear liquid height on the tray, in. or ft. Zf Average froth height on the tray, in. or ft. Greek Symbols Constant in correlation equation Gas holdup, ZfZc, sft. Pf Gas holdup in the froth, fto Gas holdup in the bubble formation zone, ft. ~P'O Constant in a correlation equation Constant in a correlation equation P tt Constant in a correlation equation Y Crozier mixing parameter 1r Liquid flow rate, lb/hr/fto E Fractional error 3i ~Humidity of the gas, lb vapor/lb dry nitrogen 1Humidity of saturated gas at Tas, lb vapor/lb dry nitrogen as e RResidence or contact time of the gas in the plate liquid X mG/L Latent heat of vaporization at the temperature Tas, BTU/lb. XTw Latent heat of vaporization at the temperature Tw, BTU/lb. 11B Solvent viscosity, cp. I-G rGas viscosity, lb mass/ft-hr. PL ~Liquid viscosity, cp or lb mass/ft-hr. Mising parameter PG Gas density, lb/ft3 xxii

NOMNCIATURE (CONl''TD) pL Liquid density, lb//ft3 PB Density of Qsolent, gm/cc. PMG Molal gas density, lb moles/ft3 Molal liquid density, lb moles/t3 XT PMean residence time of the total flow Mixing fl.w term in the differential equations describing the transfer of material to a differential element on a tray. 0'2 Variance of liquid-residence-time distribution cTL Liquid surface tension, lbft/hr2-ft. a-B Surface tension of solvent, lb:t/hr2 ft xx-.iii

INTRODUCTION The separation of materials is a very important operation in any large refinery or chemical plant. The most common type of separation means used in these plants is fractional distillation wherein the difference in volatilities of the components in a liquid or gas mixture are used advantageously to perform a separation. Another basic type of separation means is absorption where a high boiling liquid is used to absorb a particular component or components from a gas mixture. In both operations, liquid and vapor are contacted countercurrently and the separation depends upon mass transfer between these two phases. In the petroleum industry, the trend for years has been to make sharper separations of basic crude fractions and raw products from cracking operations. For example, the products from thermal cracking of light hydrocarbons and from steam cracking of gas oils and the light ends from catalytic treatment of naphthas and gas oils contain valuable chemical building-blocks such as ethylene, propylene, normal butylenes, isobutylene, butadiene, isoprene and higher boiling liquid. These hydrocarbons are usually required in high purity in the petrochemical or chemical industry. In order to recover these materials efficiently, superfractionation is required. Absorbers are also used in some cases to remove small amounts of impurities from these hydrocarbon streams. The engineer must be in a position to design the equipment required to perform these separations with a considerable degree of accuracy in order to maintain capacity and purity in the new plant and thereby maintain an attractive return on the plant investment. Fractionation and absorption are also basic tools in -1N

-2the refining of petroleum crudes and in the removing of heavy hydrocarbons from natural gas streams. However, the designs in these cases are usually based on broad experience in these operationso In addition, the designs are not as critical as in the case where a high purity prod, uct is of interest, Using the concepts of the equilibrium stage, the engineer may make a theoretical study of any separation provided vapor-liquid equilibrium data are available. With the use of digital computers, studies of this type are feasible even for multicomponent systems containing twenty to thirty componentso However, translating the results from the theoretical study to the separation equipment in the plant requires a basic knowledge of the column performance in relation to theory. The most important information required in this ease is the performance of an actual stage (bubble-cap, sieve, or perforated tray) in relation to the theoretical stage This is commonly expressed. as plate efficiency, a factor which can be related tothe mass transrier on the plate. In order to predict the plate effeciency for a particular plate design, the engineer must also be familiar with the hydraulic characteristics of the-tray. In the past, several correlations of plate efficienPy have not included factors such as liquid and gas holdup on the tray. But these factors have been shown to be very basic to any prediction of plate efficiency (7,35,36,37,92)). The use of plate efficiency is a convenience which has been accepted by design engineers. Basically, the design of a fractionator can be made by use of rate-process concepts. Stage-wise calculations could be performed on a non-equilibrium-stage basi ats oppo sed to the equilibrium stage calculations commonly used today. In this case, the

-3rate of transfer between -the gas and liquid phases would have to be known for each component: in the system. Similarly, when using the equilibrium-s tage design method and plate efficiency which is related to mass transfer on the tray, the efficiency for each component in the mixture must be known. Therefore, it i's necessary to know the effect of physical properties on the rate of mass transfer in a system of this kind. Plates are used in industry in order to obtain efficient contacting in most cases when large diameter columns are required. Where the column diameter is below two to three feet, packed columns are used. The factor commonly used to describe the performance of packed columns in relation to the theoretical stage is the-height of a theoretical plate, HETP. This'factor can also be related to the rate of mass tralofer occurring in the column. On this basis, it is not difficult to see the similarity between the two types of equipment, The principal difference is the method of contacting. In addition, it is conceivable that some of the information from the studies of mass transfer in a plate column, can be applied in the studies of packed columns or vice versa. In the following sections, the discussion will be limited to plate efficiencyo Several basically different plate efficiencies have been studied, correlated, and used by the engineer in the past. These are: I. Point efficiency, n - Yn-1 (1) n n-l

where the primes indicate vapor concentrations at discrete points lying on a vertical line through the tray, and where = vapor concentration above the tray, mol: fraction. vapor concentration below the tray, mol fraction. = concentration of vapor in equilibrium with the liquid on the tray, mol fraction. 2.- MUrphree plate efficienciess, (a) n terms -of vapor composition, Yn - Yn- (2) Yn. V Yn-l where Yn average concentration of the vapor leaving the tray, mO1 fraction. average concentration of the vapor entering the tray, m ol fraction. Ynt z= concentration of the vapor in equilibrium with the liquid leaving the- tray, mol fraction, (b) In terms of liquid composition, Xn - Xn+l Xn. Xn+l where xn = concentration of the liquid leaving the tray at the outlet downcomer, mol fraction.

-5n+1 = concentration of the liquid. entering the tray at the inlet downcomer, mol fraction. xn -= concentration of the liquid in equilibrium with the gas or vapor leaving the tray, mol fraction. For the case of straight, but not parallel, operating and equilibrium lines, EMV and EML may be related by the following material balance relationships, EMV EIEJ E-.+ mG (1- EML,) or ML ( EML where m - the slope of the vapor-liqui.d equilibrium curve of y versus x. H = Henry's Law constant, H = p/c 3. Overall Column Efficiency, number of ideal plates 0 number of actual plates where Eo = overall column efficiency. The amount of academic research work on plate efficiencies in the past is probably greater than the work on any other piece of equipment used by the chemical and petroleum industries. This work includes studies of the following types:

1. Investigations of plate efficiency wherein small columns or single trays were used to study the effect of one or two variables on the efficiency of a particular system. 2, Investigations of the hydraulic characteristics of single trays including studies Of entrainment~ flooding, and gas and liquid holdup on the trays.. 3. Investigations directed toward a better understanding of of the mechanisms involved in contacting a gas and liquid on a bubble tray. These investigations include studies of the mechanism of bubble or interfacial-area formation and the mechanism of mass transfer in each phase of the two-phase system. The petroleum and chemical industries have relied heavily on past experiences and specific tests in the pilot plant to design fractionation equipment. This.work plus the academic research on plate efficiencies add-up to a very large effort and cost for one type of equipment. In spite of this tremendous amount of effort, a general relationship between plate efficiency and. the many independent variablees is still lacking. However, the results of the previous research has shown the following variables to be important in the prediction of plate efficiency: 1. Properties of the vapor and liquid. (a) Vapor-liquid equilibrium. (b) Viscosity, surface tension, molecular weight, and density. (c) Mass diffusion rates. (d) Thermal properties, heat capacity and conductivity.

2. Equipment variables. (a) Tower diameter and plate spacing. (b) Bubble-cap design: type, size, slot dimensions. number. of slots per cap. etc. (c) Plate layout: bubble-.cap arrangement, number of bubble-caps, skirt clearance. (d) Type and height of weirs. 3. Operating variables. (a) Vapor and liquid flow rates. (b) Temperature and pressure, In view' of the number of variables involved and the number of different possible designs,. it is understandable why a general correlation of plate efficiencies does not exist today- One- of the problems which has limited the engineer in his efforts to relate plate efficiency with the fundamental variables involved has been his inability to completely describe the geometric, kinematic, and dynamic characteristics of a gas emerging from a slot and flowing through a gas-liquid mixture. Ideally, one would like to be able to describe the geometry and fluid dynamics of the system well enough in order to use the results from small scale experiments such as mass transfer to or from a gas flowing from a single slot or orifice submerged in a liquid to generalize the prediction of mass transfer on a bubble tray. Attempts have been made to do this but the relationships developed are not believed to be generally applicable. However, these studies have aided in the interpretation and correlation of the data from test trays.

-8In 1952, the American Institute of Chemical Engine.ers inaugurated a research program for the study of plate efficiency. The primary purpose of this effort has been to study efficiencies' for both small and large trays and to develop a method of predicting the efficiencies for trays of any size by use of a generalized correlation of the data from the research program. The small trays were used to study the effects of gas and liquid properties on plate efficiefncies while:the larger trays were used to study the effects of tray design variables on plate' efficiencies. An effort was made to interpret the data by use of the concepts of mass transfer in a two-phase system such.as, the two-film theory or theory of additive resistances. The present invesatigation was sponsored by the American Institute of Chemical Engineers with the primary purpose of increasing the knowledge of the -relationship between plate efficiencies and liquid properties.

TRANSFER OF MATERIAL BETWEEN PHASES Whitman(96) proposed the "two-film" theory which has been used as a model to describe the process of interphase transfer. This model implies that there is a.fluid "film" in each fluid adjoining the interface, and the principal resistance to interphase transfer is this double film. The relative resistances for the general case in a gas-liquid system depend upon two distinct sets of factors: (a) Physical factors: solubility of the gas in the liquid, diffusivity of the gas in the liquid and vapor phases, and concentration in.each phase. (b) Hydrodynamic factors: geometry and scale of equipment. viscosity and density of the two phases, and flow-rates of the two phases. If it is assumed. that the mass transfer through the adjacent vapor and liquid films takes place at steady-state and that equilibrium exists at the interface, it is possible to equate the rates of diffusion through each film to the rate through the interface. NA = kG (PG - Pi)= kL(Ci - C) KOG (PG p*) (7) where KOG = over-all mass transfer coefficient, lb. moles/(hr.) (sq.ft. ) (atm) o kG = gas phase mass transfer coefficient, lb. moles/(hr. )(sq. ft. ) (atm). kL liquid phase mass transfer coefficient, lb. moles/(hr. )(sq.ft )(lb. moles/cuft. ). PG partial pressure in gas phase, atm. -9

-10Pi = partial pressure in gas phase at the interface, atm. C = concentration in liquid phase, lb. mole/u oft. Ci = concentration in liquid phase at the interface, lb. mole/cu.ft. If Henry's Law applies Pi =HCi (8) and PAW = =HC Then the relationship between the over-all and the individual film coefficients may be obtained by eliminating Ci and pi in Equations (7) and (8). 1 _, + - H (o10 KOG kG kL 1 1 (11) KOL kL HkG In the case where the equilibrium data are in terms y-x diagrams, Equations (10) and (11) become 1 = + _ m (12) KoGP kGP kLPML and = 1 +1 (13) KLPML kLPML mkGP where m = the slope of the equilibrium curve. PML = molal density of the liquid, lb. moles/cu.ft.

Three different models for describing the mechanism-'of mass transfer in the liquid phase have been proposed. These are: 1. Whitman(96): laminar film. 2. Higbie(42): systematic surface renewal. 3. Danckwerts (22): random surface renewal. In the Whitman model the liquid at the surface is assumed to be in laminar flow parallel to the surface while the liquid below the surface is in turbulent motion. The rate of absorption is determined by molecular diffusion in the surface layers, Although the relative importance of transport by diffusion and by turbulence will vary with the depth below the surface, the model implies a completely stagnant layer of effective thickness, XL. The thickness of the film is assumed to be small enough for the absorption process to be treated as one of steady-state -diffusion through the stagnant layer. In Higbie's(42) model of systematic surface renewal, each element of liquid surface is assumed to be exposed to the gas for the same length of time and to absorb gas during this time at a changing rate as though it were a stagnant layer of infinite depth. The model proposed by Danckwerts(22) is identical to Higbiels model except for the time of exposure for each liquid element. In the Danckwerts model, it is assumed that there is no correlation between the time of exposure and the chance of an element being remixed, The physical significance given to kL is different in each of the three models as can be seen from the following expressions for the rate of absorption: Whitman: NA (C* -C) L (14)

12 Higbie: NA=( -C)2 LC (15) Danckwerts: NA = (C* (.C)2 (16) where s = the rate of surface renewal. t = contact time of a liquid element with the gas.

BASIC RELATIONSHIPS BETWEEN POINT EFFICIENCY AND MASS TRANSFER COEFFICIENTS The point efficiency is the most basic factor to consider since it does not depend on mixing or concentration gradients on the tray but only upon the hydrodynamic and physical factors which affect the mass transfer. Murphree(59) was probably the first to use the "two-film" theory proposed by Whitman(96) to derive the fundamental relation between point efficiency and the interphase mass transfer on a bubble plate. For his derivations, Murphree assumed: 1. A negligible change in total vapor volume during its passage through the liquid on a given plate. 2. Complete mixing of the liquid on the plate, i.e., no liquid concentration gradient on the tray. 3. The composition of the gas entering the plate is uniform along the length of the plate. 4, A gradual and continuous change in composition of the vapor during its upward passage through the liquid, with no back-mixing of the gas, i.e., plug flow of the gas through the liquid. Murphree's derivation is as follows: -G dy = KOG a' (PG - P*) dQ (17) -13

-14where KOGy a- = overall vapor mass transfer coefficient, mols/(unit contact time')(unit of pressture driving force)o a? = interfacial area per plate per-unit time of gas flowo:= residence or contact time of gas in the; plate liquid. G = gas or vapor rate, mols per unit timeo Since pG = Py and p* = Py - G dy = KOGa' P (y - y*) dO (18) or frQf Y * K)= aP (19) If it is assued that Ka P G and are independent of time of conIf it is assumed that K0Ga? P, G,: and y are independent 6f time of contact or in other words, independent of the position in the liquid on the tray, Equation (19) may be integrated:between the limits, y = Yl at Q = O and y = y at 0 =- 0, to obtain the following relationship between point efficiency and the mass transfer coefficient and time of gas -contact0 y -Y E = w-. = 1 - em (20) where EOG =. over-all vapor point efficiency Y1,= concentration in gas entering the tray, mol fraction of component being exchanged or transferred from gas to liquid phase or vice versa0

-15y = concentration in gas leaving the tray, mol fraction of component being exchanged or transferred from gas to liquid phase or vice versa. KOG a' P Q m = = NOG G The point efficiency is therefore defined as the ratio of the actual amount of material transferred and the amount which would be transferred if the vapor leaving the tray were in equilibrium with the liquid on the tray, Walter(89) expressed the time of contact, Q, as directly proportional to the effective slot submergence below the liquid level, h, This relation implies vertical movement and constant velocity of the gas bubble in the liquid and does not account for the gas hold-up in the liquid on the tray which increases the length of the path traveled by the gas bubble. Walter also observed that the term m, was essentially independent of the gas rate, G. In terms of Walter's definition, P KOG a' Q PKOGO a" As1 h m = -.. = NOG (21) G G where KOG a" = over-all gas film mass transfer coefficient, lb moles/(hr. )(atm) (sqoino slot area)(in. liquid depth) Asl = total slot area per plate, sqo in. h = effective liquid depth, in. (taken to be the distance from the middle of the slots to the top of the weir).

Geddes(34) attempted to study separately the factors which determine the bubble size, the time of contact, and the value of KOGo The time of contact was calculated from the liquid seal on a foam-free basis and an empirical equation for the velocity of rise of individual gas bubbles through liquids0 Chu(17), Narsimhan(60), West(93), and Colburn(20), have made similar studies where attempts have been made to separate the effects on point efficiencyo In all of these studies,: it has been recognized that the contact time of the gas with the liquid on the tray is one of the variables involved in NOG but there has been considerabble variation in the individual definition for the contact time (see Table I). Recently, Gerster(l) proposed the following definition for gas contact time,.tG =Z"c _ = (22) tu Us m = NG = K a tG (-23) where Zf observed froth height on the tray, ft. Zc = observed averige clear liquid depth on the tray, ft. Zf - Zoo effective gas holdup on the tray, ft3/ft2 of bubbling area on the tray us = superficial gas velocity through the tray, ft/sec. tG = gas contact time, sec. KtOGa = over:all coefficient, se o

17Using this definition, Gerster was able to correlate Ashby's vaporization data for several systems including water and organic liquids and gases like Freon 12, helium, air, and nitrogen. The correlation is as follows, k'Ga = C us0'23 (23a) where C = 18.19 DG0'33 DG = gas diffusivity coefficient, sq.ft./hr. ktG mass transfer coefficient for gas phase, cu. ft./(sq. ft. )( see ) a = interfacial area, sqofto/cu.ft. of gas holdup. us = superficial gas velocity, ft./sec., based on tray bubbling area. Ashby's data were obtained at one weir height, 1-1/2 inch, by e of the equipment used in the present investigation. Therefore, it is of interest to investigate the applicability of Gerster's definition as the liquid submergence on the tray and physical properties of the liquid are varied. In the derivation of Equation (20) the assumption was made that themolal rate of gas does not change as the gas flows across the tray. In the absorption of components from a dilute gas mixture, the desorption of components from a dilute liquid mixture, the vaporization of a pure liquid at low vapor pressure, or in a distillation system where the assumption of constant molal overflow applies, Equation (20) is applicable. However, in cases where the foregoing conditions do not exist, the differential equation for mass transfer at a point in the froth on a plate is

-18- G dy - ydG = K0G a P (y - y*) d3AT (24) or ady y dG KOG a P dAT (2 (25 If the change in G is caused by absorption of a single component from the gas, de.sorption of a single component from the liquid, or evaporation of a liquid by the gas, the change in G is a function of the change in the gas composition. A material balance for the inert gas (gas not transferred from gas to liquid phase) is,G - Gy = Gl - G1 Y1 (26) G = G1 (1 - yl)/( - y) (27) If G is a function of gas composition on the tray, dG d=G y- (28) dy and d.G Gl(1 ) dy (29) (1 - y)2 Substituting Equations (27) and (29) into Equation (25) dy K:G a P AT d.( (30) (Y - y*)(1 - y) G If KOG P is constant throughout the froth and if G equals Gavg Equation (3.0) can. be integrated between the limits y = Yl at 5 = 0 and y = y at, = f to give the -following relationship, N G K PjAT 1 _ * [(y* )] (51) Gavg l - y- * Y* y1 - y

-'9By introducing the molal gas density, pi, KG a P AT KG P K (2) = K0 a tG (32) Gavg Us avg PMG where KG a = over-all mass transfer coefficient, sec1. tG = gas contact time, sec.

MJRPHRBEE LIQUID POINT E'FICIENCY x ~ x1 EoL = X1 (33) Murphree introduced the liquid efficiency concept chiefly for application to the case where the mechanism of vapor-liquid interphase mass transfer makes this efficiency the more fundamental one to use, such as, a rain of liquid droplets falling through a well-mixed vapor~ In this example, the composition of the vapor in a given stage would be con — stant while the liquid. composition would undergo a gradual, continuous, change~ It is also applicable in the case of crosseflow of liquid on a bubble tray where the gas compos$ition does not change enough to significantly affect the equilibrium concentration, x * In cases where there is an appreciable change in the gas composition across the tray in the direction of gas flow an average value of the gas composition should be used in evaluating the liquid point efficiency. To be mathematically correct, this average should be determined by a double integration1 ioeo, an integration in the direction of gas flow and then an integration in the direction of liquid flow, The relationship between liquid point efficiency and the number of mass transfer units may be derived in an analogous manner to that used for vapor point efficiency0.If it is assumed that the change in total liquid flow rate is negligible, a material balance on a ddifferential element of liquid in the frQth is, -20

L x - KOL a PML(X - x*) ZATdW. L(x + dx) (34) - KOL & PM(x - x ) Z.TdW = L dx (35) Now if it is assumed that plug flow of the liquid exists and that x* and KOL a Z are constant along the length of the tray, Equation (355) may be integrated between the limits x = x1 at W = 0 and x = x at W = 1o KOL a pML Z AT 1 x (6) Lf d fW= (x56) KOL PM Z AT = fn ( x (37) L Xi.X3 L -LKOL a PML Z AT N KOL a PM: Z = -,n (1 - EO.) (38) Much like the case of NOG, the specifie interfacial area in NOL has been defined in several different ways depending upon the investi~ gator, In Table I, the definitions of N0G and N0L used by several investigators are presented. Gerster(2) has defined the specific interfacial area as square feet per cubic feet of liquid.holdup on the tray. In this case, NOL is defined as follows, KOL a pML Z AT KOLa PML Zc AT;_ __ _ __ _ = s KOL a tL (39) OL=' L L where a -= interfacial area, ft2/ft3 liquid holdup on the tray. Z~ = clear liquid height on the tray, ft.

-22The relationship between the point efficiency and the resistance in each phase may be derived as follows: y + dy Zf + dZ +ff LAafx L dZf(x + dx) Zf- Zc - kG a P(y - i) dZf(- Z ) AT dW = GdydW (40) c@ dZf (dx) kL'oa PML(xi - X) dZf dW AT z -(41) - KOG a P(y - y*) dZf Zc) dW AT = GdWdy (42) Zf If y - mx, Equations (40), (41), and (42) may be combined to obtain the following relationships, 1 __ = 1, + mG/L KOG a d(Z Z)AT kG af - Z )AT kL a P4ML d Zc AT (43) G G L 1 1 mG/L ~Kt: a dtG k'Ga dtG kL dtL (44) OG Ga G.kLa K a'Ga L -kL (45) (KOG 3 - P9ta l kf~a t (6

-231 = 1 mG/L (47) KOG a tG k'Ga te kL t tL 1 = +G/ (48) NOG NG NL

TABLE I EFIITIONS OF NOG AND NOL Investigator NOG a tG, 0 Asl Zv, h, Zc KOG Murphree(59) P KOG a G/G area per unit gas residence time lb. moles/(hr)(ft2)(atm) time of gas on plate contact Walter, et al.(89) P KOG a =Al h area per unit slot area effective liquid depth lb. moles/(hr)(atm)(ft2) G liquid depth per unit slot area West, et al.(93) KOG a Zv/G area per foot form height above plate lb. moles/(hr)(ft2)(atm) of foam Sherwood and Pigford(78) KOG a ZV/G area per foot effective depth of lb. moles/(hr)(ft2)(atm) of liquid liquid I hold-up 53 Gerster(1) - a a erster) OG = K tG area per unit = gas contact lb. moles per (hr)(ft2) gas hold-up us volume of gas us time (lb. moles/ft3) ft3 per ft2 hold-up tray area N a tL Asl Zv, h, Zc KOL Walter, et al.(89) a Asl h area per unit slot area effective liquid depth lb. moles/(hr) (ft2) L liquid depth (mol fraction) per unit slot area Sherwood and Pigford(78) KOL a PML ZV area per foot effective depth of lb. moles/(hr)(ft2) L of liquid liquid (lb. mole/ft3) hold-up Gerster(l) KOL a PML Ze AT area per cubic liquid contact time average clear liquid lb. moles per (hr)(ft2) L foot of liquid height on tray (lb. moles/ft3) hold-up

BASIC RELATIONSHIPS -BETWEEN POINT EFFICIENCY AND PLATRE EFFICIENCY The relationships between point efficiency and plate efficiency have been derived by use of the basic rate-process concepts. The liquid flows in the horizontal direction across the tray and contacts gas flowing in the vertical direction. Therefore, the concentration in the liquid changes continuously across the tray and if the gas entering the tray is uniform in concentration, the driving force for mass transfer varies along the length of the tray. Kirschbaum(50,51) Stone(85), and Peavy and Baker(66), removed liquid samples from various points on a bubble plate in relatively large columns and found appreciable differences in concentration. Gerster(2) and Warzel(92) found significant concentration gradients on trays of medium size. Gadwa(31) took liquid samples from a five-inch tower and fcund small horizontal liquid concentration gradients but concluded that their effect was negligible. In order to predict plate efficiency starting from a general correlation of point efficiency, it is necessary to be able to predict the variation of the liquid concentration with distance along the trayo This problem has been studied by several investigators (4,21 30 32 50o 51 7285) and the solution to the problem has varied considerably, depending upon the method or model used to characterize the liquid mixing on the tray. The differential equations or difference equations and the solutions to these equations which have been used by various investigators to describe the system of mass transfer- and liquid mixing on bubble trays are presented in Table IIo The necessity for knowing the amount of liquid mixing can be best explained by pesenting the derivations for some of the relationships between point efficiency and plate efficiency, -25

TABLE II RELATIONSHIPS BETWEEN POINT AND PLATE EFFICIENCY Terms in Differential Equation Interphase Investigator Flow Mixing (j - r') Transfer Boundary Conditions Relationship Between Point and Plate Efficiency Lewis(53) L d Plug flow = 0 EOG(mx - Yl) x = at W = 1 EM/Eo = (eXG )/E XEOG Warzel(92) L dx (C - 1)L dx/dW EOG(mx - yl) x = xO at W = 1 EMV/EoG = (e C 1)/ G XEOG Crozier(21) L d[xn(W)] Kd[x (W+AW)] EOG(mx - yl) x = xO at W = 1 EMV/E = (el+y - l)/ XE dW - DE Ze B pML ex = Xo at W = i RobinsDn(72) L dx/dW - DE Zc B PML d2x EG(mx - ) dx/dW at W = EMV/EG dW2 {dW =0 at W = I (I+M)(Pr+M tit]+M Robinson(72) L dx/dW - DE Zc B PML d2x E X- at W = 1 ~S d2dX O t Wl E~~ _[1(d= x at W = 1i)(2 Anderson(4) L dx/dW - DE Zc B PML d2x EOG(mx Yl) x = xO at W = 1 Xl Xe DE Zc B PML dx at W - EML= (2n+M)eTi+M 2+ (T + ZL dWx x -ZL dW

-27A material balance for a differential element in the froth on a bubble tray is shown in Figure (1) for the material being transferred from the gas to the liquid phase or vice versa. Inlet Downcome Outlet GY1dW Figure'l Mixing Model Used to Derive Relationships Between Point and Plate Efficiency. The material transferred to and from the elements by liquid mixing on the tray is represented by the terms, tr and t' For the case of plug flow of the liquid, material balance on the element is, Lx + GyldW = L(x + d dW) + GydW (49) or L w - G(y - y1) (50)

-28By use of the definition for point efficiency EOG.= (y- yl)/(mx - Yl), the term, (y - Yl), may be replaced by EOG (mx - Yl) to give the following result: dx mG GEOG (51) dW L L The general solution to the above first-order differential equation is, x -yly/m =ctL-EOGW (52) if it is assumed that EOG is independent of the position on the tray Other assumptions which are implied in the derivation of this equation are: (1) the gas entering the tray is homogenous in concentration, (2) the number of moles of gas remains constant as it rises through the liquid, and (3) gas distribution is uniform over the plate. If the boundary condition of x = xo at W = 1 is applied, the relationship which may be used to predict the concentration at any point on the tray is, x - yl/m _ eEo(Gl - W) (53) xo - Yl/m where X = mG/L In order to arrive at the relationship between point and plate efficiency, Equation (53) must be combined with the definitions for EOG, Murphree point efficiency, and Emv, Murphree plate efficiencyo FOG - (54) mx - y]_

-29or y = EOG mx - EOG Yl + Y1 (55) Yavg - Yl EMV = mxo - Y1 where Yavg = the average concentration of the gas leaving the tray. mxo = y, the gas in equilibrium with the liquid leaving the tray. If the following definition for Yavg is used, 1 fO ydW = ydW (57) Yavg~ ~ foYavg = 1 EOG mxdW - E yjdW + yldW (58) Substituting in Equation (58) the relationship for x as a function of W, Yavg = jo EOG m(xo - Yl/m) eXEOG(1 - W) dW + jo yldW (59) or Yavg = EOG m(xo - Yl/m)(e EOG 1) + Yl (60) XEOG Yavg - YlE (eXEoG_ 1) (61) mxo - Y1 XEOG e-EOG- 1 EMV/EOG XEOG (62) XEOG Equation (62) is the relationship derived by Lewis(53) for the case where liquid flows across the tray without mixing and the vapor entering the tray is uniform in concentration. Lewis developed equations for two other types of vapor-liquid contacting but these are not thought to be applicable in most bubble plate columns.

In. -the cases where the m.ix.ng terms, i and 4ft, in Figure 1. are not zero (i.e., when the assumption of plug flow does not apply)! these terms'must appear in the material balance on the differential element used to derive the differential equation which describes the system. Warzel(92) defined a mixing parameter "CO in terms of the terminal concentrations on a bubble-cap tray identical in design to the one used in the present investigation. Warzel's definition of "Co" is, C =x o (63) Xe - x where xo = concentration in liquid leaving the tray, mol fraction. x1 = concentration in liquid entering the tray, mol fraction. Xe = concentration in liquid at a point on the tray between the inlet down comer and the first row of bubble caps. According to Warzel's definition, "C"s approaches one as plug flow is approached and approaches infinity as complete mixing is approached. The material balance on the differential element for the case where the mixing parameter "C" is used is as follows: CLx + GyldW = CL(x + d+ dW) + GydW (644) The differential equation for this case is, CL d - G(y - Y1) (65) dW Thus, according to Warzel's definition, the mixing on the tray is characterized by an effective flow rate, (C-l)L. The solution to Equation (65) is similar to the equation for the case of plug flow.

A C (1 - W) x - y1/j=m _ (iw) (66) o - l/m The relationship between point efficiency, EQG, and plate efficiency, Emv, is obtained by using Equation (66) and the procedure used for the case of plug flow. XEOG EMV e C - 1 _-=....(67) EoG XEOG/C Crozier(21) used a differential difference equation to'.derive a relationship between point efficiency and plate efficiency using a mixing parameter very similar: to that used by Warzel. Crozier defined the mixing parameter in terms of the terminal concentrations on the bubble tray as follows, X1'- Xo x1 X - X0 X (68) Xe - Xo Xe where the nomenclature is identical to that used in the:definition of the parameter, "C". Using the above definition, Crozier derived the following relationship, -EOG 3EM (.e0 +Y _ 1) (69) EOG XEOG/(1_*) Another approach to the problem of mixing has been through the use of the concept of ideal mixing stages. Kirschbaum(50,51) was probably the first to report the use of ideal mixing stages in relating plate and point efficiency~ Others who have studied the problem of using ideal mixing stages are Nord(61), Gautreaux and Q'Connell(32), and.Foss(30).

-32 Gautreaux and O'Connell(32) were able to derive the following analytical expression to relate point and plate efficiency which is much simpler than the suggested solution by Kirschbaum. (51) EMV _ [ G X)n (7 ) EOG XENG n They postulated that "n",, the number of ideal mixing stages, is a function of the length of liquid flow path and liquid and gas rates. According to Kirschbaum(51) and Gantreaux and O'Connell(32), a tray with an infinite number of ideal mixing stages corresponds to the case of plug flow, and a tray with one mixing stage: cOrrespohds to a:perfectly mixed tray. A tray with the number of ideal mixing stages between one and infinity represents the case of partial or incomplete mixingO A third method of characterizing the mixing on bubble trays has been the use of eddy diffusivity where the mixing terms, j and l' in Figure 1 are determinedtby the concentration gradient in the froth. The material balance on the differential element in Figure 1 for this case is as f.ollows-, Lx + GyldW. Zc B ~dx - L(x + dx dW) + -GydW S dW dW EZ B ML (dx d( )dW S dW dW d ZC B A4L d2x Ldx G(y - Y1) (72) S dW2 dW'

-33Replacing y - yl by E0G (mx - Y) Zc B L d2x Ldx- _ G EOG mx EOG Gyl(73) S dW2 dW d2x LS dx G SEoGmx G S EoG (74) ze B PML dW E ZcW B PL DE Zc B pL The general solution to this second-order differential equation is, x - ylm =C e(fl+M)W + c2 e-lW (75) where l= _- M +i +X EOGM 2 4 M ZL DE tL i Z, B PML DE tL The constants, C1 and C2, in Equation -(75) are determined by the boundary conditions at the terminal points of the tray. Gerster and Robinson(35) applied the boundary conditions, x = x0 at W 1. dr _ 0 at W = 1 dW and obtained the following relationship between liquid concentration and the fraction of distance along the tray. x - Yl/m _l e(rM)(W-l)e+ (1 ) e (76) Xo - Yl/m 2+M 2q+M When this equation and the equation which defines EOG are combined and the integration along the length of the tray is performed, the following

relationship between E0GJ and EMV is obtained. M_. = 1 - e-(~+M) eq ~ 1 (77) EOG (I+M) God) (2 T) Anderson(4) applied the same set of boundary conditions to obtain a solution very similar to Equation (77) by Robinson (72) How ever, Anderson used the boundary condition at the inlet to the trays, i.e0, xl = Xe -DE_ Z'c B pl dx (78) X1 X e, (78) SL.dW to obtain the following relationship-* xl -y/m_ [l1 - r(1 )] -(+M) xo y/m 2 ~q+M) + ( _ -.)(1 +.) el (79) - 2 q+M M Now if yl/m = x X1'- Y1/mxl- x _ EML (80o) xo- ym xo - x 1 - E Equation (8o) is valid Qnly in the case where the change in concentration of the gas does not change significantly as it flows across the tray, i,e,s when Y1 = Yavg above the tray. Foss(30) has shown the plate efficiency to be a function of the residence-time distribution of the liquid on the tray. The rate of mix — ing was characterized by Foss(30) by the rate-of-increase-of-variance of the ages of fluid elements as they flowed across the tray, The expressions derived by Foss included a distribution function, f(t). of the liquid residue time on the tray.

-35E 1 - e-XEOGt/Tf(t)dt e f(t)at (81) EOG ( XEOGA e=XEOGt/ f(t)dt where T = mean residence time of the total flow. t - residence time of a differential fluid stream. The result for the Murphree liquid efficiency is, -l1[l- -0 e-f EOGt/Tf(t)dtJ The integrals in Equations (81) and (82) are the La Place transformation of f(t) and it is possible to use a table of transforms to compute EMv/EoG and EML quickly when a functional form of f(t) is given. Foss showed that the rate-of-change-of-variance determined during his investigation could be correlated with the froth momentum as follows, d2 = A (z V)-2.8 (85) = ( f(83) dW Ze Zf where Zc/Zf = froth density. By -use of the relationships between eddy diffusion coefficient and the residence-time distribution developed by Danckwerts(23), Foss also showed that the rate-of -change-of-variance for the liquid-residencetime distribution could be related to the diffusion coefficient by the f ollowing equation, dcx2 2 D (84) dW Vf where a 2 = variance of liquid-residence-time distribution. Vf =mean linear froth velocity, ft/sec.

-36 In addition, Foss(30) used Whartons( 95) eddy diffusion data for a seive trays 1-foot wide and 5-feet long- and Equation (84) to calculate the rate-of-change-of-vari ance for the seive tray and showed that the resulting data were correlated very well by use of Equation (83). Thus, Equation (84) seems to be the link between the eddy diffusion coefficient and the residence-time distribution. The eddy diffusion coefficients reported by.Wharton(95) were determined by injecting a salt solution into the froth and then measuring the concentration in the froth at several points upstream from the injection.: The gradient on the tray was used to determine the eddy diffusion coefficient by use of the following equation, log C f + K (85) 2DE which is the solution to the differential equation for unidirectional flow. Stone(85), Robinson(72) and Brown(9) have obtained similar data for bubble-cap trays.

PREVIOUS INVESTIGATIONS OF THE EFFECT OF LIQUID PROPERTIES-VISCOSITY AND DENSITY In Table III, some of the systems studied in the academic field of research are tabulated. In addition, the type of equipment used, range of efficiencies obtained, and the estimated range of liquid density and viscosity covered by the systems are included in Table III. One possible explanation for the limited data on the effect of liquid viscosity on plate efficiency is the fact that the greatest interest has been in the field of fractionation. The liquids are in many cases at their boiling point in fractionation and therefore only small changes in liquid viscosity are encountered. This follows to a certain extent from the rule-of-thumb that liquids at their normal boiling points exhibit about the same viscosity. The highest liquid viscosity used to-date in a study of plate efficiency is about 22 centipoise. This work was done by Walter 89) The equipment consisted of a 2-inch diameter column with a segment of a 2-inch diameter cap. Propylene and isobutylene were absorbed in a virgin heavy naphtha, viscosity 1.04 centipoises at 77~F; a virgin gas oil, viscosity 6.2 centipoises at 77~F; and a mixture of a virgin gas oil and an SoAoE. 30,"Zerice" lubricating oil, viscosity 23 centipoises at 77~F. The data from this study plus the data of Horton(44) and Fairbrother(28) for the desorption of carbon dioxide from water and from glycerol-water solutions were used to calculate mass transfer coefficients by use of the relationship suggested by Murphree (59) EMV = 1 - exp(Ka As * h) (86) -37

-38where KGa = over-all gas film coefficient, lb. moles/ (hro)(sq.ino slot area)(in liquid depth) (mole fraction driving force). As, = total slot area per plate, sq. in. h = effective liquid depth, in (taken to be the distance from the middle of the slots to the top of the weir or in some cases, the differences between the dydrostatic head and the height of the middle of the slots above the plate. G = gas rate, lb. moles/hr. Values of the single film coefficients were then obtained by-use of the following relationship: /KGa - l/kGa+ (87) HkLa The results for kLa at a constant slot velocity were correlated by the following equation: kLa = 34/iL o58 (88) In order to develop a relationship between.all of the variables involved in their study and the over-all mass transfer coefficient,. Walter and Sherwood(89) used the data of Carey eto al.(ll) for the rectification of ethanol-water to estimate the effect of slot velocity. The over-all coefficient was shown to be- proportional to the cube root' of the slotwidth. The effect of liquid visecosity on the gas-phase mass transfer coefficient was assumed to be identical to that found.for the liquid coefficient (see Equation 88).Experimental data were used to show that the

-39TABLE III SUMMARY OF PREVIOUS EFFICIENCY STUDIES Liquid Properties Investigator Distillation Systems Column-Size Density Viscosity Pressure, ATM Range of Efficiency, % Gadwa(31) Alcohols-water, benzene-carbon 5-inch square EMV, 70 to 99 tetrachloride Langdon and Keyes(52) Isopropanol-water EMV, 70 to 92 Lewis and $oley(53) Benzene-toluene-xylene EMV, 30 to 75 and Nord( 1 Carey, Griswold, Lewis Benzene-toluene 8-inch and 6-inch EMV, 70 and McAdams 11) diameter Ethanol-water 6-inch diameter EMV, 50 to 100 (single plate) Aniline-water 8-ipch diameter EMV, 58 (ten plates) Uchida and Matsumoto(87) Ethanol-water, methanol-water 25-cm diameter EMV, 85 (41 plates) EMV, 70 Rhodes and Slachman(71) Ethanol-water, benzene-toluene EMV, 80 EMV, 65 Peavy and Baker(66) Ethanolwater 18-inch diameter (10, 3-inch caps 3 plates) GrJhse, et al.(31) Extractive distillation 15-inch diameter EMV, 47 to 61 of C4 hydrocarbons (ten plates, 13 Caps per plate) Schilling, Peyer, and Ethanol-water 18-inch diameter EMV, 100 WatsonI75) (3 plates, 10; 3- EOG, 83 inch caps per tray) Oliver and Watson(64) Acetone-water, ethanol,water, 18-inch diameter EMV, 40 to 100 ethylene dichloride -toluene (3 plates, 10, 3- EMV, 95 inch caps per tray) Mayfield, et al.(56) Propanol-sec-butanol EMV, 63 Rush and Stirba(73) Acetic acid-water, methyl 18-inch diameter isobutyl ketone-water (sieve tray) Humidification Systems Walter(89) Humidification of air 2-inch diameter 1 EMV = EOG, 85 to 90 (segment of a 2inch cap) Farrell and Vyverberg(29) Humidification of air 2-inch diameter 1 - 4 EMV = G (segment of a 2-inch cap) Polich(6 Humidification of air l0-inch diameter 1 EMV = EOG, 68 to 92 (4, 2.4-inch caps) Gerster, et a1.(35) Humidification of air 15-inch diameter ESV = EG, 87 to 98 (13, 1.5-inch caps) West, et al.(93) Humidification of air 3-1/4 x 2 inchesMV o E1G 76 to 95 (sieve tray 83, 1/8-inch diameter holes) Ashby(7) Adiabatic vaporization of 7-1/2 x 11-13/16 0.6 2.86 EMV EO, 78 to 96 water and several organic inches (9, 1-1/2liquids inch caps) Absorption Systems Walter(89) Propylene and isobutylene 2-inch diameter 1 - 20 cp 3.15 - 4.5 EML, 12.5 to 65 petroleum fractions (segment of a 2-inch cap) Horton(44) Carbon dioxide in water 18-inch diameter 1.1 - 1.3 cp 1 EML, 60 to 90 (7, 4-inch caps) Fairbrother(28) Desorption of carbon dioxide 5-inch square 0,5 - 3.7 op EML, 35 to 81 from glycerol-water solutions (one 3-1/2 inch and water cap) Etherington(27) Absorption of propane in 2-inch diameter 0.4 - 5.0 cp 5.4 heavy naphtha containing (segment of a vistanex 2-inch cap) Stone(85) Gerster, et al.(35) Desorption of oxygen from water 9.5 x 45 Inches 1 EM, 25 to 80 Polich(68) Desorption of oxygen from water 10-inch diameter 1 EML, 45 to 92 (4, 2.4-inch caps) West, et al.(94) Desorption of carbon dioxide 4 x 3-1/4 inches 1 EL, 67 to 85 from water (sieve plate 83, 1/8-inch holes) Warzel(92) Absorption and desorption, 7-1/2 x 11-13/16 1 *I 40 to 90 carbon dioxide-water inches (9, 1-1/2inch caps) * Where not otherwise indicated, density and viscosity of the systems are estimated to be in the ranges of 0.5 - 1.0 gm/cc and 0.5 - 1.5 cp.

-40single phase coefficients were proportional to V/Asgo These results and assumptions were then used with Equation (87) to develop the following correlation: E (1 - e-m) (89) where = h (90) (2 5o + 0 ~370/P)LO~ 058DsO 33 Etherington(27) studied the effect of viscosity upon the mass transfer on a bubble cap tray by absorbing propane in a single oil and varying the viscosity of the oil by adding small amounts of "Vistanex", a high molecular weight polymer of isobutylene. The small 2-inch diameter column used by Walter was also used by Etherington. Runs were made at oil viscosities of 0.9, 7,9, and 16 centipoises. Pure propane was used as the gas in order to eliminate the effect of the gas phase resistance. Values of the liquid-film -coefficient were calculated by use of the equations used by Walter and Sherwood. (90) (See Equations 86 and 87)o Etherington used the following relationship to correlate his data plus the data by Walter(90), Horton(44), and Fairbrother(28): (PKLa"')(C2m/2o) [ L(s + Ds) ]0 52 ( (91) 1o67 0~67 0078 T (DL) (PL) where B and A are functions of Liquid viscosity and where C2m = the average concentration of the solvent in the liquid film, lb. mols/cuoft. C20 molal density of the liquid phase9 lb. moles/cu.ft.

This relationship was derived by use of the following basis: 1. The equivalent of the Chilton-Colburn(l5) evaluation of the total effective liquid-film thickness 2. Geddes (34) method of evaluating gas bubble diameter, velocity, and flow path in the bubble plate liquid. 3. Arnold's(5) empirical equation for liquid diffusivity. 4. Sherwood's(76) definition of liquid diffusivity and the liquid-film mass transfer coefficient. In effect, Etherington's results indicate that the dependence of the liquid phase coefficient on liquid viscosity varies with the liquid viscosity. Etherington attributed this effect to a transition from film resistance to eddy resistance controlling liquid-phase mass transfer, or, in other words, a transition from laminar to turbulent film resistance in the liquid. In the range of viscosity between 0.4 and 3.0 centipoises, Etherington reported the value of A in Equation (91) to be 0.17 and between 4.0 and 20.0 centipoises the value was 0.73. Between 3.0 and 4.0 centipoises a transition region was indicated. During the course of Ashby's studies(7) on the effect of gas physical properties upon mass transfer in the gas phase, the liquid viscosity was varied between 0.5 and 2.5 centipoise. The data from this study have been correlated satisfactorily without including liquid viscosity in the correlation.(l) Quigley, Johnson and Harris(70) studied the effect of liquid properties, viscosity and density, upon the gas holdup due to air flowing through square-edged orifices submerged in various liquids. The liquid density was varied from 62.4 to 98.0 lb./cu. ft. and viscosity from 1.0 to 400 centipoises. Holdup was found to be a

function of the air flow rate only. H -2.44 x 10-4 QG 0o84 (92) where H = holdup of gas in froth, cu~uft./cuofto of froth. = volumetric air flow rate per orifice, Cu. ft /hroa Average bubble sizes were determined bY a stroboscopic light technique. TheI effect of liqulid,viscsity -on the bubble diameter was found to be small as indicated by the following equation, db = 0.222 do ~033 QG 0, 125 JL/PL 00- (93) where.db = diameter of equivalent spherical bubble,, ft. do = Orifice diameter, frt I/P.L = kinematic liquid viscosity, sqofto/hro Smolin(81) studied the hydraulics for air water on a tray very similar to the one used in the present.investigation. The tray was 10 by 12 inches and contained 1-1/2 inch diameter caps. The arrangement of the caps and the skirt clearance were varied. The viscosity of the water was varied over a narrow (up to 5 cp) by the addition of glyeerolo Although the cap arrangement on the tray and the cap skirt clearance affected the gas and liquid hold.-up on the thray, no effect due to change in.liquid viscosity was notedo Studies directed toward determining the effect of liquid density on plate efficiency are noticeably lackingo -Efficiency data are available for systems which cover a narrow range of densities, water and hydroearbons, but no,attempt has been made to correlate these -data directly~ oa-bn, bu -n.t re

Geddes(34) arbitrarily assumed that the bubble formation could be described by equating the buoyant force to the surface tension force at incipient break away of the bubble from the slots. The diameter of the bubble in this case is inversely proportional to the liquid density to the one-third power and the interfacial area is inversely proportional to liquid density to the two-thirds power. Experimental studies of bubble formation at orifices and slots have shown that the assumption made by Geddes is valid only at very low gas rate. Silberman(74) reports that there are actually three regimes of bubble formation depending on the gas rate through the orifice or sloto These are: (1) The single bubble regime, bubble size dependent on surface tension, liquid density, and orifice diameter and independent of gas rateo (2) An intermediate regime, bubble size and production rate increase with increase in gas-flow rate. (3) Jet regime, gas emerges in a more or less continuous jet which breaks up outside the orifice to form bubbles. The rate of bubble formation is nearly constant and the size increases with gas-flow rate0 Correlations of over-all tower efficiency for numerous commercialsize fractionators with liquid viscosity have been presented by DCrickamer and Bradford(26) and O'Connell.(62) Drickamer and Bradford's correlation is as follows: Eo= 0.18- 0.60 logl0LI (94) where pIL = molal average viscosity of the feed at the average column temperature, centipoises. EQ = over-all column effiiency.

This correlation was improved by O'Connell by including relative volatilityo Chu(18) has recently proposed the addition of two more factors, liquid to vapor mass ratio. and submergence, to improve the correlationo,

PURPOSE OF PRESENT INVESTIGATION The primary purpose of the presext investigation is to decrease the uncertainties in accounting for the effects of_ the liquid properties, viscosity and density, on the mass transfer in the two-phase system on a bubble plate. Although the effect of liquid viscosity has been studied over a significant range in previous investigations, several factors have not been investigated thoroughly during these studies. The most important of these are: 1i When the viscosity or density of the liquid.on a bubble plate is varied it is very likely that the hydraulic characteristics of the trayl or more directly, the gas and liquid holdup on the tray, vary accordingly. Now if the results of the mass transfer studies are interpreted without accounting for these variations in gas and liquid. holdup as liquid viscosity is varied, nothing has been gained toward. determining how viscosity affects plate efficiency. It is of value.to determine whether a change in viscosity affects the interfacial area for mass transfer, the mechanism by which mass is transported in the individual phases, or both of these. 2. The effect of diffusivity on mass transfer in the liquid phase has been taken.into account in previous studies by use of the dimensionless number,,L the Schmidt s PL Number. In most cases, the liquid properties, viscosity and density, were not varied over a significant range to

-46justify the use of the,Schmidt numbero In fact, the liquid diffusivity -could have been used to give an equally good correlation, The penetration theory proposed by Higbie(42) and Danckwerts(22) predicts' that the mass transfer -coeffi — cient is proportional to' the liquid diffusivity-.to the one — half power. In the present'investigation an attempt wasmade to determine whether the Schmidt number or -liquid diuffusivity is needed in the correlation of mass transfer coefficients for bubble-cap trays. 3. Data.on the effect of'iiquid-viscosity on mass transfer in the gas phase have not befen reported. This is especially true for the case where the viscosity has been varied over a significant range. A liquid with a high viscosity and.adequate vapor pressure was used in the present study to determine this effect by adiabatic' vaporization of the liquid. Efficiency data have been obtained using various liquids of different densities~ However, the range of densities covered is approximately 0.5 to 1O0 gm/cc and no significant effect'of liquid density has been reported, If the Schmidt number is used to correlate the data for mass transfer in the liquiid phase,- liquid density is automatically included in the correlation with samte power on liquid density and diffusivity. Of course, this combination does not necessarily have to correlate the data. Some investigators(l8s34L) i attempts:to correct Tfor liquid density have assumed that bu'bles are produced at the slots, one at a time, with the size of each determined primarily by the orifice diameter, surfaee tension,

-47and buoyancy in order to develop a relationship describing the interfacial surface area on a bubble tray~ In this case the interfacial area is inversely proportional to the liquid density to the two-third power and independent of the gas rate. However, studies of the bubble size formed as gas flows through slots and orifices submerged in liquid indicate that the above mechanism does not apply since bubble size has been found to be a function of the gas rate. Although the variation in liquid density will not be as great as the variation in liquid viscosities in most fractionators or absorbers, it is conceivable that a range of 0.2 to 30 gm/cc would not be out of the question. Ethylene dibromide with a density of about 2.2 gm/cc was used to extend the range of liquid density used in the investigation of mass transfer in the gas phase.

SUMMARY OF EXPERIMENTAL -INVESTIGATIONS The experimental studies in this investigation are divided into -the following three categories: (1) The mass transfer resistance in the gas phase on.a bubble tray where liquid viscosity and density are varied. (2) Mass transfer resistance in the. liquid phase where the liquid viscosity is varied. (3) Studies of the hydraulics, gas and liquid holdup, on a bubble tray where the liquid viscosity and density are varied. These studies were performed by use of the followiing techniques: (1) Gas phase resistance, -(a) Adiabatic vaporization of cyclohexanol; gas phase, nitrogen; liquid viscosity varied by varying liquid temperature (b) Adiabatic vaporization of ethylene dibromide; gas phase, nitrogen; liquid density about 2~2 gm/cc. (2) Liquid phase resistance, (a) Absorption of carbon dioxide in cyclohexanol; liquid viscosity varied by varying.liquid temperature0 (b) Liquid concentrations at four points on the bubblecap tray were used to correct plate efficiency to point efficiency~ -48

-49(3) Hydraulic characteristics, gas and liquid holdup, (a) Average liquid holdup was determined by measuring the hydrostatic head of liquid at four different points on the tray. (b) Gas holdup was determined by subtracting the average liquid holdup from the average froth height, The average froth height was determined visually by use of a graduated scale on the glass window which covered one side of the tray.

MATERIALS Nitrogen and Carbon Dioxide Nitrogen and carbon dioxide in cylinders were obtained from the Liquid Carbonics Corporation by the General Stores at the University of Michigan and was supplied for use in the present investigation. Cyclohexanol (C6Hll0H) The cyclohexanol was supplied gratis by the Dow Chemical Company, Midland, Michigan~ Some of the properties of the cyclohexanol were given by DowO (25) These are presented in Table IVo TABLE IV PHYSICAL PROPERTIES OF DOW CYCLOHEXANOL(25) Freezing Point*, OC Boiling Range, 5-95% at 760 mm Hgt, C 156-163 Specific Gravity, 25/25~C 0: 948 lb/gal at 25 C 7.89 Refractive Index at 25~C 1l.469 Flash Point, F -145 * Sets to a glasslike solid, no freezing point, ** Literature value for pure cyclohexanol, 23o9OCo(67) The impurities reported by Dean(24) are phenol, maximum, 0o5 weight percent; cyclohexanone, maximum, Oo.l weight percent; and water, maximum, 0o5 weight percento -50

Ethylene Dibromide (CH2BrCH2Br) The ethylene dibromide was also supplied gratis by the Dow Chemical Company. The properties of ethylene dibromide reported by Dow(25) are presented in Table V. TABLE V PHYSICAL PROPERTIES OF DOW ETHYLENE DIBROMIDE(25) Freezing Point, 0C 9,3 Boiling Point, ~C at 760 mm Hg F13104 Specific Gravity, 25/250C 2.172 lb/gal at 25~C 18.07.Refractive Index at 250~C 1o536 The data in Table V are approximately the same values reported in the literature(67) indicating that the ethylene dibromide supplied by Dow does not contain significant amounts of impurities.

EQUIPMENT AND LABORATORY PROCEDURES A detailed description of the equipment and laboratory procedures is presented in Appendix A, B, and C. In this section, a bribef description of the equipment and a summary of the laboratory procedures are presented. Equipment The equipment used in this study was also used by Warzel(92) and Ashby. (7) Certain modifications of the equipment were made in order to perform the studies in the present investigation. The modifications are listed in Appendix A. Figure 2 is a general view of the test column plus the auxiliary equipment. The bottom tray in the column.was used for a test tray and the-second tray from the bottom.servedas an..entrainment separator for the first tray. The dimensions of the tray and the tray layout are presented in Table VI. Briefly, the tray was 7-1/2 inches wide and 11-13/16 inches long and contained 9, 1-1/2 inch diameter, caps..A close-up view of the test:tray equipped with caps and the adjustable splash baffle is shown in Figure 3. The outlet weir on the tray was adjustable and studies were made at weir heights of 1-1/2, 2, and 3-1/2 inches. In order to eliminate a large hydraulic gradient on the tray, a splash baffle near the outlet weir was used. This baffle was positioned one-half inch.above the top of the weir and one inch in front of the weir in all studies performed. A glass panel in the front of the test tray was used to permit observation of the hydraulic characteristics of the tray.

Figure 2. Front View of Test Column and Auxiliary! Equipment.

-54TABLE VI CHARACTERI STICS OF THE TRAY AYOUT -..... I. i Cap Diameter (O.D.) -ll/2 inch Height 1-l/2 inch Metal Thickness 1/16 inch Slot Height 3/)4 inch Width 1/8 inch Number, per cap 18 Area, per cap 0.0117 sq. ft Area, per plate 0-.105 sq. ft. Area, fraction of bubbling area 0.171 Risers Diameter (O.D.) 1 inch Diameter (i.Do ) 7/8 inch Area, per cap 0.00Q17 sq. ft. Area, per plate 0o03,75 sq. lftn Area- fraction. of bubbling asea 0 061 Weir (variable) Length 7-3/8 ihes Height ll/2, 2, and 3-1/2 inches Splash Baffle (variable) Length 7-3/ iAces Clearance above tray floor 2,2-1/2 and 4 inches Downcomer Size 7-3/8 x 2-/8 inches Area, cross sectional:0.37 sq. ft. Bubbling Area (taken as space between downcomer and splash baffle) Width 7.1/2 inches Length 11-13/16 inches Area 0.615 sq. ft. Tray Spacing 18 inches

-55-.....~ B i lief~~~~~...... ] i _~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~: Figure 3.Single Tray in Absorber., Showing Removable Deck and Adjustable Splash Baf fle. i i i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i'Figure 3. Single Tray in Absorber, Showing Removable Deck and Adjustable Splash Baffle.

-56A rotary blower (two rotary blowers in the absorption studies) was used to recirculate the gas to the test tray. A Durco pump equipped with a mechanical seal was used to recirculate liquid to the test tray. In the vaporization studies, a second Durco pump was used to recirculate the same liquid as used in the tests to a 12-inch diameter, 4-plate sieve cq1ump where the gas from the test tray was dehumidified by countercurrent contact at a lower-temerature than the temperature on the tray. In the absorption studies, one of the pumps was used to feed liquid from the sieve column toithe test tray and the other one was used to take liquid from the test tray and return it to the sieve-tray column. The sievetray column in this case was used to desorb carbon dioxide from the liquid.with air from the laboratory. The air to the sieve-tray column was supplied by a third rotary blower. The liquid and gas to the test tray were metered by -use of calibrated Fischer-Porter rotameters. Calibration data for these rotameters are presented in Appendix D. The temperatures of the gas and liquid at the test tray were controlled by.use- of heat exchangers in the gas and liquid lines. Calibrated mercury-in-glass thermometers were used to measure the temperatures at the test tray. The pressure in the system was controlled by addition of nitrogen in the vaporization studies and by addition of carbon dioxide in the case of the absorption studies. The rate of addition was controlled by a "Pancake" pressure regulator. The pressure.above the test tray was measured by a mercury manometer,

-57Laboratory Procedures - Vaporization Studies In the vaporization studies, the mass transfer on the tray was determined by measuring the concentration of the gas entering and leaving the tray. These concentrations were determined experimentally by sampling the inlet and outlet gas after adiabatic conditions were established on the tray. The humidity charts in Appendix B were used to calculate the inlet gas temperature when the humidity of the inlet gas and the liquid temperature or adiabatic saturation temperature of the gas were known. The vapor in the gas was determined by passing gas samples from the test column through a series of three U-tubes filled.with glass beads and immersed in ice baths. The amount of organic vapors condensed in the U-tubes was determined by weighing the tubes before and after sampling. The volume of each sample was determined by metering the gas in a wettest meter after leaving the series of U-tubes. The inlet and outlet samples were taken simultaneously while the temperatures on the tray were maintained at adiabatic conditions. The data recorded. for each run are shown in the sample data sheet in Appendix F.Laboratory Procedures - Absorption Studies The change in the concentration of carbon dioxide in the gas across the test tray was not great enough to be determined accurately, Therefore, liquid samples at the inlet and outlet of the tray were taken and. analyzed to determine the amount of mass transfer on the tray. In.addition, samples of the liquidd.at four different points on the floor of the tray were taken and analyzed in order to correct the over-all plate efficiency to a point efficiency. The concentrations of the liquid in the froth were determined at three different horizontal positions for

-58one series of runs, The locations of the liquid sampling points on the test tray are shown in Figure.4. The position of the sample points on the tray floor in relation to the position of the caps is shown in Figures 5 and 6 The liquid samples were obtained by use of either a hypodermic syringe or a valve connected to the sample point by a 1/4 inch stainless steel tube. When the hypodermic syringe was being used to take samples the desired volume of liquid was withdrawn and injected into the sample bottle which contained 50 ml of 0o 1Ba(OH)2 and was sealed with a syringe stopper~ However, when the samples were being taken through the 1/4 inch line and valve, the liquid flowed directly into the sample bottle and a magnetic stirrer was used to mix the cyclohexanol and barium hydroxide. The two methods of sampling were cqmpared by taking consecutive samples from the same point and analyzing for carbon dioxide~ At the higher concentrations (at tray outlet) the results-agreed within one percent while at the lower concentrations the disagreement was as much as five:percent. The amount of carbon dioxide in the samples was determined by titration of the excess barium hydroxide in the sample bottles by use of O,1 N'HC1. Phenolphthalien was used for an end-point indicator. A sample of the gas leaving the tray was obtained.during each run and analyzed for carbon dioxide in order- to' determine the equilibrium concentration on the tray. The analyses of the gas samples were made by use of a CEC mass spectrometer, Type 21-103Co Laboratory Procedures - Hydraulic Data The hydraulic data obtained during this study consist of the average froth height, the clear height or hydrostatic liquid head at four different points on the trays, and the pressure drop in the gas flowing

-59FLOW 21" 0X 9 0 0 0 / ~~I | >vSAMPLING POSITIONS ABOVE TRAY FLOORR / 21/2'" 25/16-K2 3/84 3 3/4 SAMPLING POSITIONS IN FLOOR OF TRAY *..... 1/2" 7 "DOWNCOMER SPLA$S BAFFL FROTH LIQUID SAMPLING _ _.. l/$" TUBING *1 PROBES 3/8" TUBING I" LONG OVERFLOW N2,,: NO.6 AT SOTTOM OF OOWNCO~Rt IQUID SUMPS USED TO 0 6' Figure 4. Location of Sampling Points on Test Tray. Points 2,3,4 and 5 Used Also to Measure Hydrostatic Head at Respective Points

-60Figure 5. Top View of Removable Tray, Showing Bubble Caps and Location of Liquid Sampling Points. Figure 6. Bottom View of RemovaDle Tray, Showing Inlet to Risers and Liquid Sumps Below Sampling Point.

-61across the tray. These data were recorded during the course of the studies of vaporization and absorption. The measurement of the froth height was done by visually observing the average height of the froth above the tray floor, Two scales with Oo1 inch graduations placed near the edges of the bubbling area but on the glass panel covering the front of the tray were used to determine the average height. The variations in froth height across the tray plus the fluctuations due to instability of the system were taken into account when determining the froth height.. The clear liquid heights were measured at the four points indicated in Figure 4. The pressure drop across the tray was measured by use of a water manometer attached to pressure probes above and below the tray.o

PRESENTATION OF DATA.Hydraulic Data - General Observations The general observations of the hydraulic characteristics which were made during the mass transfer studies and which are worth consideration in the interpretation of these data pertain to the effect of the following variables: 1o Gas rate. 2. Liquid rate 3. Weir- height. 4. Liquid viscosity and density. Some of the effects of varying liquid and gas rate at a constant weir height and constant liquid viscosity and density can be seen by examining the photographs in Figures 7 and 8. These photographs were made by enlarging single frames from movies made at a film speed of 64 frames per second. In Figure 7, the general appearance of the froth holdup on the tray for the carbon dioxide-cyclohexanol system is shown for the cases of F-factor, 0.62; liquid viscosity, 55 centipoises; and weir height, 3-1/2 inches, at liquid rates of 26.4, 16.2, and 6.6 gpm. The detail in the photographs is not great enough to show the bubbles, or maybe better described as globules, in the froth However, the photographs do show clearly the effect of the entering liquid on the froth holdup. At the liquid rate of 6.6 gpm, the three rows of caps on the tray were active and the froth was fairly uniform along the length of the t. But as liquid ray.te was increased to 12 gm the slots on the side of the first row of caps nearest the inlet doncorer became inactive and a large eddy began to develop in the volume of froth between the first row of caps and the inlet downcrner, At a liquid rate of 26.4

liquid rate, 26.4 gpm F-factor, 0.62 liquid, rate, l3.2 gpm F-factor, 0.62 i —-y ~~~~ —~liquid rate, 6.6 gpm Figure 7. Froth Holdup on Test Tray, Carbon-Dioxide-Cyclohexalol System; Weir Height, 3-1/2 Inches; Splash Baffle Height, 4 Inches F-Factor, 0.62; Liquid Viscosity, Approximately 55 cp.

-65gpm, the liquid eddy extended almost to the center of the tray and the gas ensuing from the caps passed through the froth at an angle to the vertical. At a higher gas rate, the entrance e ffects of the liquid were not as pronounced. In Figure 8, the froth holdup on this tray is shown for the cases of F-factor, 1.1, and liquid viscosity, 55 centipoises, at liquid rates of 26.4, 16.2 and 6,6 gpm. In these cases the slots on the side of the first row of caps nearest the inlet downcomer were completely active or intermittently active. Also the gas seemed to rise almost vertically through froth at a liquid rate of 26.2 gpm and not so much at an angle to the vertical as in the case of a lower gas rate of F-factor, 0.62. It therefore appears that the liquid entrance effects disappear at gas rates above F-factor equals one, However, these effects are significant at high liquid rates and at gas rates below F-factor equals -one. Another observation which can be made by examining the photographs in Figures 7 and 8 concerns the amount of froth holdup in the outlet downcomer as the liquid and gas rates are varied. As might be expected, the holdup in the outlet downcomer increased as the liquid or gas rate increased. Photographs or movies of the froth holdup on the tray at lower weir heights were not obtained. However, the liquid entrance effects were observed to be similar to those shown in Figures 7 and 8. The principal difference between the characteristics of the froth at 1-1/2 and 2 inch weir heights and those at 3-1/2 inch weir was the form in which the gas flowed through the liquid. At 1-1/2 inch and 2 inch weir heights, there were two distinct types of bubbling. At low gas rates, the gas

-66flowed through the froth in form of globules or bubbleso At high gas rates, the globules seem to disappear and the gas flowed through the froth in channels or sheets, The value of the gas rate at which the globules disappeared and the gas began to flow through the froth in channels was a function of weir height. With the weir height at 1-1/2 inches the gas globules were broken-up at gas rates above Q08 F-factor while with the weir height at 3-1/2 inches this did not occur at all up-to an F-factor of 1.2. The effects of the liquid properties, viscosity and density, on the visual appearance of the froth holdup were very noticeable when the froth holdup for the carbon dioxide-cyclohexanol and nitrogenethylene dibromide systems were compared. In the case of the cyclohexanol, at all viscosities, 10-100 centipoises, the gas flowed through the froth in the form of large globules or channels depending on the velocity and weir height with veryfew smaller bubbles or globules formed due to break-up,of the larger gas volumes~ The froth holdup for the ethylene dibromide system contained many small gas bubbles giving it a very uniform appearance. However, the small bubbles masked the action of the gas flowing from the caps and it was impossible to see the size of gas partidles in that areas Hydaulic Data - Gas and Liquid Hoidup The clear liquid height or hydrostatic head of liquid was measured at four different points on the tray floor (see Figure 4). The arithmetic average of these four measurements was used as the average clear liquid height or the liquid holdup on the tray. The froth height was determined by visually averaging the variations in froth height along

the length of the tray and the time fluctuations due to the instability of the system. These data were obtained during the course of the mass transfer studies, The gas holdup was determined by subtracting the average clear liquid height from the froth height~ In Figure 9, the clear liquid height on the tray is presented for the nitrogen-cyclohexanol system as a function of the superficial velocity in the column for variable conditions of weir height, liquid viscosity, and liquid rate~ Similar data for the nitrogen-ethylene dibromide system are presented in Figure 10. The effect of increasing the weir height is about what would be predicted, i.e., one-half inch or.two inch increase in weir height increases clear liquid height about the same amount. This, is not exactly true as can be seen in Figure 9 for the nitrogen-cyclohexanol where an increase of one-half inch and two inches in the weir height increased the liquid holdup by 04 - 005 inch and 1o0 - 1.4 inches, respectively, It should be noted that in no case does the clear liquid height equal the weir height in Figure 9 except in the case of data for the 3-1/2 inch weir where the clear liquid height decreased to a value below the weir height as the, gas velocity was increased.. For the nitrogen-ethylene dibromide system, the liquid holdup for 3-1/2 inch weir and 7.9 gpm was equal to the weir height and independent of gas velocity. In all cases in Figures 9 and 10, the liquid holdup is not highly.dependent on gas velocity. The effect of increasing the liquid rate is shown by data forthe 2 inch weir in Figure 9 and for the 3-1/2 inch weir in Figure LO, In each case the liquid holdup increased as the liquid rate; was increased indicating an increased liquid head required to overcome the resistance

5.0 4.0 w z 3.0 IA 0~~~~~~~~~~~~~~~~~~ U 2.Z 60 m I =~~~~~~~~~~~~~6809M12 cp 2 W.EIR 0 8.0 gpm /L= 25 P 2 WEIR22ofWE -1 WEI 16.0 9Pm I'LL2 OP 0 1.0 0 8.0 9PM4LL25p 22 5 CP 12 p 3 WEIR A 8.0 9pm L~L1 p 0,0 4,O 5.4 0 2.0 1.0~ ~~~~~~~~~~~. SUPERFICIAL VELOCITYll ) ft/sec Figure 9. Clear-Liquid-Height Data for Nitrogen-Cyclohexanol System

5.0 A A~~~~~~~~~~ 4.0 U) wr I5J A~~~~~~~~~~~~~~ U ~~~~~~~~~~~~~~~~A 3.0 w ~ 2.0 I~WEIR A 15.9gqpm, 32 WEIR 011 0.5 1.0 2.0 3.0 4.0 5.0 SUPERFICIAL GAS VELOCITY, UsJft/sec. Figure.0. Clear-Liquid-Height Data for Nitrogen-Ethylene Dibrominde System

5.0 4.0 Cl) w c 3.0... Cy 2.0 O (D 0 0 A A - O O 000 00 WWj 0 4.94 GPM,F&Lu- 94.0-98.5 cp 0' 1.0 4) 4.90 GPM,'L' 23.5-24.6 cp A 16.5 GPMqUL'23.8-24.5 cp 0 26.5 GPM, ILL 23.8'-24.6 cp p 0 0.2 0.4 0.6- 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 F-FACTOR, us A Figure 11. Clear-Liquid-Height Data for- Carbon dioxide-Cyclohexanol System at High Gas Rates and Two-Inch Weir Height.

-71to flow at the splash baffle and weir. Liquid viscosity effects are not considered significant since the data for the two systems in Figure 9 and 10 are about the same at comparable weir heights and liquid rates. The only effect of viscosity is the possible difference in the dependence of liquid holdup on gas velocity. In the case of the cyclohexanol system, the liquid holdup at 3-1/2 inch weir height decreased as the gas rate was increased but the data for the ethylene dibromide system was practically independent of gas velocity at the same conditions. In Figure 11, the clear liquid height for the carbon dioxide-: eyclohexanol system is presented as a function of gas velocity with variable conditions of liquid.rate and viscosityy The range of gas velocity in this case is almost twice that for the data in Figures 9 and 1Q0. Froth height data for the three systems, nitrogen-cyclohexanol, nitrogen-ethylene dibromide, and carbon dioxide-cyclohexanol are presented in Figures 12, 13, and 14 as a function of gas velocity with parameters of liquid rate, weir height, and liquid viscosity. The effect of weir height and liquid rate on froth height is very similar to the effect on clear liquid height. That is, as weir height or liquid rate are increased froth height increases proportionately. However, froth height is definitely dependent on gas velocity in contrast to the almost complete independence of clear liquid height and gas velocity. Liquid viscosity or density does not appear to affect froth height significantly evenpthough the viscosity is varied over the range of 10.5 to 25 centipoises and density over the range of 0.94 to 2.15 gm/cc.

10.0 8.0 (I)~~~~~~~~~~~~~0 IZ a.- 6.0 I 4.0 1 1 4.0'~~~~~O 8.0 gpm,'LLL 25 cp I WEIR n ILI E 8-0 25cp, 2 WEIR PL~~~~~~~~~~~~~~~~~~~~~ 2.0 ~ 8.0 9pm,/LL 12 cp) 2 WEIR Z16.0 gpm).L: 25 cp, 2 WEIR A 8.0 gpM: 12 cp, 3- WEIR 0 1.0 2.0 3.0 4.0 5.0 SUPERFICIAL VELOCITY U ft/sec Figure 12. Froth-Height Data for Nitrogen-Cyclohexanol System. Data Are Presented as a Function.-of Superficial Gas Velocity With, Parameters of Weir Height, Liquid Viscos ity and Liquid Rate.

12.0.... 6I.0 UA Z I. 4' 8.0 I- I 6.0 — I0.5 1.0 2.0 3.0 4.0 SUPERFICIAL VELOCITY, t t./sec. Figure 13. Froth-Height Data for Nitrogen-Ethylene Dibromide System. Data are Presented as Rate Ra-t e

14.0 13.0 12.0 A- O 11.0 10.0 0 w 9.0 —. 08.0 CD 7.0 A 6.0 0,c 5.0 to ~ ~~L~~~~~ O 4.94 GPM, pl=94.0-98.5 cp 4.0- 0 4.90 GPM, IL=23.5-24.6 cp A 16.5 GPM, ILL23.8-24.5 cp 3.0 0 26.5 GPM, /L=23.8-24.6 cp 2.0 1.0 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 24 2.6 F- FACTOR,1JS/'pG Figure 14. Froth Height Versus F-Factor-Carbon Dioxide-Cyclohexanol System Variable Liquid Viscosity 2-Inch Weir; 21-Inch Splash Baffle.

-75In Figures 15 and 16 the froth height data obtained by Ashby(7) plus the data for the nitrogen-cyclohexanol and nitrogen-ethylene dibromide systems are plotted versus gas velocity and F-factor. In Figure 15, where the data are plotted versus gas velocity the best curves through the individual sets of data do not coincide. However, when the same data.are plotted. versus.F-factor, the data may be represented by one curve. Therefore, it appears that the major variables at constant liquid rate and. weir height are gas velocity and gas density and not liquid viscosity and density, It should be noted that the ranges of liquid density and viscosity represented by the systems in Figures 15 and 16 are 0.78 to 2.15 gm/cc and 0.5 to 24 centipoises, respectively~ Gas holdup was determined by subtracting the average clear liquid. height from the froth height, These data for the nitrogencyclohexanol and nitrogen-ethylene dibromide systems are plotted in Figures 17 and 18. Of course, the effects of weir height and liquid rate are similar to the effects of these variables on clear liquid height and froth height. In addition, comparison of the data in the two figures reveals no great effect of liquid viscosity and density except at 3-1/2 inch weir height and high gas velocity. This is shown more effectively in Figure 19 where the data are plotted versus F-factor. Ashby's data are also included in Figure 19. The data for 1-1/2 inch weir height are represented by one curve with no significant deviations due to liquid properties. However, the data for 3-1/2 inch weir height at comparable liquid rates can not be represented by one curve. The gas holdup for the nitrogen-ethylene dibromide system is greater than that for the cyclohexanol and water systems with the lowest values for

-768.0 7.0 0 6.0 0 (I) w(: 5.0 4 44.0 3: p =0.070 lb/ft3' N2-CzH4Br2, I/L = 1.45cp PL2.1 5gm/cc PexO.070 lb/ft3 0 N-'CYCLOHEXANOL,L L 24.Ocp PL=O.94 gm/cc PG=O.010 Ib/ft3 A He-WATER I, L 0.77cp,PL=0.99 gm/cc PG 0.068 Ib/ft3 A AIR-WATER,C/L -0.77cp, PL=0.99 gm/cc 2.0 Peo 0.27 lb/ft3 V FREON 12-WATER, LL 0.60CP, PL 0.99gm/cc pe=0.013 Ib/ft3 X He-iC4H9OH IJL -2.3Ocp, PtL0.78 gm/cc p6=0.016 Ib/ft3 V He-MIBK,LL 0.50 cp, PL3.078 gm/cc 1.0 PG 0.066 lb/ft3 * N2-iC4H90H IL =2.30cp, PL =0.78 gm/cc. I 1.. I I-I 0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 SUPERFICIAL VELOCITY, U5, ft/sec Figure 15. Froth-Height Data for Several Systems Used in Vaporization Studies. Data are'Presented as a Function of Superficial Gas Velocity With Parameters of Liquid Viscosity and Density and Gas Density; 1-1/2 inch Weir Height; 8.0 GPM

-778.0 7.0 _)D XQ 6.0 _ x~ =5.0 X N z 4.0 S A* 0 3.0 A N2-C2H4Br2,, Ls 1.45 cp PL 2.15 gm/cc O) N2-CYCLOHEXANOL,'LL 24.0 cp PL 0.94 gm/cc A He-WATER /L 0' 77 cp' PL 0.99 gm/cc 2.0 AIR-WATER,, L= 0.77 cp PL 0.99 gm/cc O FREON 12 - WATER. L a 0.60 CP PL' 0 99 gm/cc X He - i C4H9OH,/L 2.30 CP,,L 0. 78 gm/cc + He- MIBK /'LO. O cp P PL 0.78 gm/cc 1.0 N2- IC4HOH,'aLL 2.30CP P*. 78 gm/cc 0 I 0.4 0. 0. I!. I 0 0.2 0.4 0.6 0.8 1.0 1.2 F-FACTOR, Us VPG Figure 16. Froth-Height Data for Several Systems Used in Vaporization Studies. Data are Presented as a Function of F-Factor with Parameters of Liquid Viscosity and Density; 1-1/2 inch Weir Height; 8.0 GPM.

7.0 o 8.0 gpm,L=25cpI, I WEIR 0 8.0 gpm, IL= 25 cp, 2 WEIR 6.0 * 8.0 gpm, LL= I 2 cp, 2 WEIR | 16.0 gpm,,UL=25 cp, 2' WEIR wtt 5.0 A 8.0 gpm, 2= cp,3 WEIR I" z 2.0.5 4.0 2.0 3.0 4.0 SUPERFICIAL VELOCITYus, fsec Liquid Viscosity, and Liquid Rate. 1.5 1.0 2.0 3.0 4.0 5.0 SUPERFICIAL VELOCITY, US, ft/sec Figure 17. Gas Hold-Up Data for Nitrogen Cyclohexanol System. Data Are Presented as a Function of Superficial Gas Velocity With Parameters of Weir Height, Liquid Viscosity, and Liquid Rate.

12.0 0 11/2-INCH WEIR,7.9GPM A 31/2- INCH WEIR,7.9GPM A 3 1/2-INCH WEIR,15.9GPM 10 0 8.0 C-) z ~ 6.0 0 0., 4.0 0 1 1.0 2.0 3.0 4.0 SUPERFICIAL VELOCITY, ls,ft/sec Figure 18. Gas Hold-Up Data for Nitrogen-Ethylene Dibromide System. Data Presented as a Function of Superficial Gas Velocity With Parameters of Weir Height and Liquid Rate.

-808.0 7.06.0 0 3- INCH WEIR ~I ttINCH WE Nr 4.0 0~ ~0 a (cp) pL(gm/cc] A N2- C2H4Br2 1.45 2.15 ai N2 - CYCLOHEXANOL 12.0 0.94 2.0 N2- CYCLOHEXANOL 24.0 0.94 A3 0 He- WATER 0.77 0.99 O AIR-WATER 0.77 0.99 o FREON 12-WATER 0.60 0.99 1.0 O N -i C4HOH 2.30 0.78 He - i HgOH 2.30 0.7q o~a 0.6 0.6 o.8 1.0 12 F -FACTOR, us/I Figure 19. Gas Holdup Data for Several Systems. Data are Presented as a Function of F-Factor with Parameters of Liquid Viscosity and Density and Weir Height; 8.0 GPM.

A81the air-water system. Therefore, it appears that gas holdup at weir heights above 1-1/2 inch increases as either liquid viscosity or density is increased. Additional evidences of a liquid viscosity; effect is indicated in Figures 20, 21, 22, and 23 where the gas holdup data for the carbon dioxide-cyclohexanol system and Warzelis data(92) for the airwater system are compared at equivalent weir heights and liquid ratesh In all cases except in Figure 23 the gas holdup for the cyclohexanol system is higher or equal to the gas holdup for the air-water system. The data for the air-water system in Figure 23 were obtained at a slightly higher liquid rate than used in obtaining the data for the carbon dioxidecyclohexanol system. This might account for the reverse relationship between gas holdup and viscosity. The difference between the gas holdup for the air-water systems and the cyclohexanol systems is, in most cases, no greater than one inch. In fact, the average deviation is probably nearer one-half inch. The significance of these deviations is questionable since the data are from two independent investigators and the greatest deviations were found at the higher gas velocities where the visual measurement of froth height is more difficult because of the rapid time fluctuations in the froth. Hydraulic Data - Relative, Froth Density Relative froth densities were calculated by dividing the clear liquid height in the middle of the tray by the froth height. If there were no significant stratification of the froth, this ratio would be a good estimation of the relative froth density. Crozier(21) used a light transmission technique to determine the froth stratification on a tray identical to the one used in the present investigation. Considerable

0.7 8.0 0.6- T. 0 6.0 — 0.5 5.00- 0.4 — 2: w w IL. z 4.0 -- N N3.F 0 -A- CO2 - CYCLOHEXANOL 0.2 26.4 gpm,.L - 24 cp, 2" WEIR 0.2 — U) 2.0-' 0 AIR- WATER 27.5 gpm,.LL: 0.58 cp, 2" WEIR O.I0 0.0 0.2 0.4 0.6 0.8 1.o 1.2 1.4 1.6 1.8 F -FACTOR Figure 20. Gas Holdup Data for Carbon Dioxide-Cyclohexanol and Air-Water Systems as a Function of F-Factor.

6.0 0.5 E3 5.0 0.4 Iw Q 4.0 w U.D Ct o.3 C.) N z I' 3.0 — N O.2 N 4 2. 4.95Cgpm,&= 52 cp,3 WEIR 0 CO - CYCLOHEXANOL 0.1AI -W 1.0 ElAR-AE N J21" 4.95 gpm //L 52 c., 3 WE IR u) Q C02 F- FACTOREXANO L Figure 2].. Gas Holdup Data for Carbon Dioxide.-Cyclohexanol and Air-Water Systems as a Function of F-Factor. I~~~~~~~~~~~~~~~~~~W I; 4.95 9pm /.LL=4 cp, 3 WEI 1. 0 1 [3 AIR- WATER /~ i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~II' 4.6 gpm /.IL= 0.6cp, 3 WEI.0 ~I 0.0f... 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 F - FACT OR Figure 21. Gas Holdup Data for Carbon Dloxlde-Cyclohexanol and Air-Water Syrstemzs as a F'unction of F-Factor.

0.7 8.0 0.6- 7.0 6.0 0.5 w 5.0 1f 0.4 0 Z N -4.0 N 0.3 I -J - 3.0 3.0 3I~~ 7:B CO2o - CYCLOHEXANOL 0.2 16.2 gpm /L= 52 cp, 3 WEIR 2.0 0E AIR- WATER 13.73 gpm /L =0.6cp, 3o WEIR ~~~0.1 - ~ 0 AIR-WATER 1.0 i"//1 / 18.31gpm /.L =0.6cp, 3$ WEIR 0, 0.0 1, 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 F - FACTOR Figure 22. Gas Holdup Data for Carbon Dioxide-Cyclohexanol and Air-Water Systems as a Function of F-Factor.

0.6 7.07 1 /t l El 6.0 0.5 5.0 - IL l 1~ N u 4.0 0.3.0.2 a 30 A CaO -CYCLOHEXANOL 0 2,.. z: 25.8 gpm, 0L 6 52 cp, 3i WEIR N 0.2 1.0 0. 0' I I I I I.0.0 0.2 0.4 0.6 0.8 1.0 1.2.4 F- FACTOR Figure 23. Gas Holdup Data for Carbon Dioxide-Cyclohexanol and Air-Water Systems as a Function of F-Factor.

-86variation of the froth density in the vertical direction was found for the air-water system at gas rates up-to a F-factor of 1.2 and liquid rates from 5 to 32 gpmo At low gas rates, this stratificatiion can be seen by simply observing the froth holdup on the tray. The appearance of the froth for the organic systems and the air-water system were compared by this technique. The froth stratification for the cyclohexanol systems was believed to be greater than that for air-water system. One possible explanation for this difference is the higher liquid viscosity which caused the gas to flow through the froth in large globules without breaking-up to form a more uniform froth densityo The froth for the nitrogen-ethylene dibromide system contained many small bubbles and was more uniform in appearance than that for either the- cyclohexanol or water systems. With considerable stratification in the froth, the ratio of the clear liquid height and froth height has less meaning as froth,density. However, this ratio might be very useful in describing the hydraulic characteristics of the tray. In Figure 24, the relative froth density for the systems studied in this investigation plus the data of Ashby(7) and Warzel(92) are plotted versus F-factor, These data are for weir heights of 1-1/2, 2, and 3-1/2 inches and represent a considerable range for liquid rate and liquid properties. Although the maximum deviation from a single curve through the data would be about 0.1, no definite trend with liquid properties is noted. This is shown more effectively in Figures 25, 26, and 27 where the density data for each weir height are plotted versus F-factor. The average deviatliorn from a single curve through the data is smaller than the average deviation in Figure 24,o IHowever, the

1.0 1-1/2- INCH WEIR 2- INCH WEIR 0.9. He - H20, 8.0 GPM @ C02 -CYCLOHEXANOL, 4.88 GPM, /LL=97.3 cp O AIR- H20, 8.0 GPM CO2-CYCLOHEXANOL, 4.85 GPM, JL =24.2 cp \ FREON 12-H2 0, 8.0 GPM { C02 -CYCLOHEXANOL,16.5 GPM,,L = 24.2 cp UI He -i C4H9 OH, 8.0 GPM 0 N2-CYCLOHEXANOL, 8.0 GPM,L =12 cp 0.8d \( N2-i C4H9OH, 8.0 GPM Nz -CYCLOHEXANOL, 8.0 GPM,/L - 25cp N2- C2 H4 Br2,7.9 GPM AIR-WATER, 4.58 GPM \ + \I RN2 - CYCLOHEXANOL, 8.0 GPM, /L =25 cp AIR-WATER, 27.47 GPM,* 3" AIR-WATER, 41.20 GPM 0.7 - \ NOTE: Zc =AVERAGE OF TWO POINTS IN THE CAP REGION 3-1/2-INGH WEIR N 0.6- O 002 -CYCLOHEXANOL, 4.95 GPM,L = 53.4 cp | CG 02 -CYCLOHEXANOL, 4.95 GPM,/LL = 24.6 cp A C02 -CYCLOHEXANOL, 16.2 GPM,,/L = 52.5 cp (n, 4 CO2 - CYCLOHEXANOL, 25.8 GPM, PL =48.7 cp 0.5 N2- CYCLOHEXANOL, 8.0 GPM,/L =12.0 cp ~~~~o 0 5 \4n(>t >\~~~~~ *~ N2 -C2 H4 Br2,7.9 GPM "'r, AIR -WATER, 4.58 GPM: 0 \ Ir3> < + AIR -WATER, 9.16 GPM 0.4 * AIR - WATER,18.31 GPM a AIR - WATER,27.47 GPM A AIR - WATER,41.20 GPM:: 0.3 ". 0. I 0.1 I I I I I I } I I I I,' 1 l 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 F- FACTOR, Us'PG Figure 24. Relative Froth Density For Several Systems. Data Presented as a Function of F-Factor with Parameters of Weir Height, Liquid Rate, and Liquid Properties.

-881.0 0.9 \ He-H20,8.0 GPM \ AIR-H20, 8.0 GPM _8 \ - to FREON 12-WATER, 8.0 GPM 0. 8 U | He-i C4Hg OH, 8.0 GPM ( N2-iC4 HOH, 8.0 GPM 0 N2- CYCLOHEXANOL, 8.0 GPM, /L = 25 c p 0.7 _- \ +- N2- C2H4 Br2,7.9 GPM \ NOTE: Zc=AVERAGE OF TWO POINTS IN 0.6 \ THE CAP REGION. 0.6 >0.5 0 w 0.4 0.3: a 0.2 0. I o0I I I I I I I I I 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 F- FACTOR, Us J/PG Figure 25. Relative Froth Density for Several Systems. Data are For l —Inch Weir and Constant Liquid Rate of About 8.0 GPM.

I.O 0.9 9 C02-CYCLOHEXANOL, 4.88 GPM,/zL = 97.3 cp C02-CYCLOHEXANOL, 4.85 GPM,,L =24.2cp 0 C02-CYCLOHEXANOL, 16.5 GPM,,L =24.2 cp U.8 C02 -CYCLOHEXANOL, 26.5 GPM.,/L =24.2 cp 0 N2-CYCLOHEXANOL, 8.0 GPM, UL= 12 cp \ N2-CYCLOHEXANOL, 8.0 GPM,pL =25cp A N2-CYCLOHEXANOL, AIR-WATER, 4.58 GPM AIR-WATER, 27.47 GPM *NYINC AIR-WATER, 41.2 GPM *: o AIR-WATER, 36.62 GPM >- 0,6 ~ O AIR-WATER, 45.8 GPM z NOTE: Zc= AVERAGE OF TWO POINTS IN THE 1 CAP REGION. = 0.5 0 0.3 0 20. 0 0. O 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 -8: 2.0` 2 2.4 -FACTOR,w Figure 26. Relative Froth Density for Several Systems. Dat4 are For 2-Inch Weir, Variable Liquid Sate and Liquid Properties.

1.0 0.9 C02-CYCLOHEXANOL, 4.95 GPMIL053.4 cp 4 C02-CYCLOHEXANOL, 4.95 GPM,pez24.6 cp COg)- CYCLOHEXANOL, 16.2 GPM,,U S525 2cp 0.8..4. C02 - CYCLOHEXANOL, 25.8 GPM,PL:48.7cp A AIR-WATER, 4.58 GPM 4 AIR-WATER, 9.16 GPM * AIR-WATER, 18.31 GPM 0.7 A A AIR-WATER, 27.47 GPM A AIR-WATER, 41.20 GPM 4 N2-Cp H4 Br2, 7.9 GPM.0j - N2NC-2H4Br2, 15.9 GPM 0.6 Np -CYCLOHEXANOL, 8.0 GPM, /PL 12cp Z NOTE: Zc=AVERAGE OF TWO POINTS IN THE o CAP REGION o 0.5 U.o~ 0.4 04 0.3A 0.2 0.1 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 F-FACTOR, US'/P Figure 27. Relative Froth Density for Several Systems. Data are For 31-.Inch Weir, Variable Liquid Rate and Liquid Properties.

-9'1 deviations are considerable and are a function of liquid rate and not liquid properties. In each of the Figures 25, 26, and 27, the data for the low liquid rates fall below the curves while the data for the high liquid rates fall above the curves, Therefore, a complete correlation of these data could be accomplished by including liquid rate and weir height. However, as illustrated in Figure 28, this correlation could not be used to predict the relative froth density for a large tray where a splash baffle is not used., In Figure 28, the data for the carbon dioxide-cyclohexanol and air-water systems at 3-1/2 inch weir height and variable liquid rate are compared -with data obtained by Gerster (2) for a larger.tray without a splash baffle but with a tray layout similar -to the one used. in the present investigation. The effect of the splash baffle is analogous to the effect of increased liquid rate since the splash baffle presents an added resistance to liquid flow and a greater liquid head on the tray is required to overcome this resistance and maintain the rate of liquid flow.* A similar effect would be expected if the outlet downcomer were not designed to handle the flow from the tray. There was some evidence that this might be the case for the tray used in this -investigation. At the higher liquid and gas rates the outlet downcomer became filled with froth which sometimes extended above the top of the weir-~ For these reasons, the hydraulic data should not be used to predict hydraulic characte-ristics of a larger tray. However, the observations regarding liquid properties. should be applicable to the larger trays. The clear liquid height is therefore equal to or greater than the liquid height for the tray with a weir and no splash baffle.

1.0 N SYMBOL SYSTEM LIQUID RATE WEIR HEIGHT RE-FEu ~ GPM/FT. WEIR INCHES RENCE * CYCLOHEXANOL-CO2 8.05 3 Ir~~~~ 1~~~~A 16.10 - 0.8 +O ~W M 24.15 = | \ x 32.20 0 0 AIR- WATER 7.45 (92) IC~~~~~~~~~~~~~~~~L ~14.90 0.6 I.6- 0 22.35 + 37.20 _I_ a ACETONE-BENZENE 30.00 4 (2), w= x t * V oII I 50.00 o 0.4O * AIR-AMMONIA-WATER 30.00 3 (2) 0 30.00: |: r* TRAY DESIGN 4 - 0. TRAY DESIGN 2 I 0.2 IL Ocr 0 0.5 1.0 1.5 2.0 2.5 3.0 F= p Figure 28. Comparison of Relative Froth Density on the Tray Used in the Present Investigation and the Density Reported by Gerster(2) for a Similar Tray with No Splash Baffle

Mass Transfer Data - Vaporization Studies The- two systems used in the vaporization studies were- nitrogencyclohexanol and nitrogen-ethylene dibromideo The conditions used in these studies are summarized as follows: Io Nitrogen-cyclohexanol (a) 1-1/2 inch weir height; ~L 25 cp; liquid rate, 8~0 gpm. (b) 2 inch weir height; L9 25 and 12 cp; liquid rate, 8o0 and 16o0 gpm. (c) 3-l/2 inch weir height; AL, 12 cp; liquid rate, 80o gpm0 IIo Nitrogen-ethylene dibromide (a) -1/2 inch weir height; liquid rate, 7o9 gpmo (b) 3-1/2 inch weir height; liquid rate, 7o9 gpm and 15o9 gpmo The range of superficial gas velocity in each of these studies was 1l0 to 4o5 ft/seco A sample data sheet and calculation rocedure for the vaporization data are presented in Appendix F. The detailed data are presented in Table I-Go The efficiency data for the nitrogen-cyclohexanol system are plotted versus superficial velocity in Figure 29.o There are three important points in regards to these data which are worth discussion0 These are: (1) decrease in plate efficiency when weir height was increased; (2) increase in efficiency when liquid viscosity was increased; and (3) increase in eff~iciency with increased gas velocity~ The decrease in plate efficiency with inereased weir height was not expected since the froth holdup or gas holdup on the tray increased

90 O 8.0 gpm, 2L 25 cp, i~ WEIR El 8.0 gpm, LL 25 cp, 2WEIR 80 8.0 gpm, iLL- 12 cp, 2" WEIR. 16.0 gpm,.L 25 cp, 2" WEIR E A 8.0 gpm, /L 12 cp, 3 WEIR W 70~ ~ 50 / I 40 I I 1.0 2.0 3.0 4.0 5.0 SUPERFICIAL VELOCITY, ls, ft/sec Figure 29. Gas-Phase Efficiencies for Nitrogen-Cyclohexanol System. Data are Presented as a Function of Superficial Gas Velocity with Parameters of Weir Height, Liquid Rate, and Liquid Viscosity

-95and the gas contact time increased accordingly. Assuming that the mass transfer coefficient, kGa, is independent of the froth height, the efficiency should have increased. Therefore, the results in Figure 29 indicate that the mass transfer coefficient is not independent of the weir height or the liquid holdup on the tray but decreases as liquid holdup increases. In fact, the decrease in the mass transfer coefficient is so great that the effect is not counterbalanced by the increase in gas contact time with increased weir height. The effect of liquid viscosity is shown in Figure 29 by the efficiency data for a 2 inch weir height and viscosities of 12 and 25 centipoises. The efficiency data at 12 centipoises are lower than the data at 25 centipoises at all gas velocities with the possible exception of the data at the gas velocity of 1 ft/sec. Thus these data suggest that theldependence of efficiency on gas velocity is a function of liquid viscosity. This variation of the relationship between efficiency and gas velocity with liquid properties is more apparent when the efficiency data in Figure 30 for the nitrogen-ethylene dibromide system are compared with the data for the nitrogen-cyclohexanol system. The viscosity for the ethylene dibromide was 1.45 centipoises and the density was 2.15, The efficiency for the ethylene dibromide system is almost independent of gas velocity. The change in efficiency over the complete velocity range is not greater than 4 efficiency percent in contrast to the change of 15 - 30 efficiency percent for the cyclohexanol system over the same velocity range. Ashby's data(7) for several systems having low liquid viscosities were also practically independent of velocity.

A 60 w I O ~ _ I \: I'Z O 7.9 gpm, AVG. GAS TEMP.: 127.7FF, 1 WEIR A 7.9 gpm,AVG. GAS TEMP 119.3 F, 32 WEIR 0 I" 0 I0 * 7.9 gpm, AVG. GAS TEMP. = 115.5 F I WEIR A 15.9gpm,AVG. GAS TEMP 119.3F, 32 WEIR 0.5 1.0 2.0 3.0 4.0 5.0 SUPERFICIAL VELOCITY, 1S, ft/sec Figure 30. Gas-Phase Efficiencies for Nitrogen-Ethylene Dibrnmide System. Data are Presented as a Function of Superficial Gas Velocity with Parameters of Weir Height and Liquid Rate

-97Another comparison which can be made between the data from the present study and Ashby's data concerns the level of the efficiency values. The lowest efficiency value reported by Ashby was 78 percent for the helium-isobutyl alcohol system. The data for the other systems were above this value with highest value of 97 percent for the Freon 12-water system. In contrast, the efficiency values for the nitrogen-cyclohexanol and nitrogen-ethylene dibromide systems were as low as 50 percent. In the studies with the nitrogen-ethylene dibromide system several runs were made before it was realized that an error had been made inadvertently in the humidity chart for this system. These data are indicated in Figure 30 by the open circles and are included herein to give some idea of the errors in the data due to the possibility of non-adiabatic conditions in some cases. A decrease of about 12~F in the average gas temperature decreased the resulting efficiency by 7 efficiency percent in some cases. The possibility of the adiabatic conditions being in error by the amount indicated above is very unlikely since in every run the gas temperature was adjusted until the inlet and outlet liquid temperatures were within 0.1 to 0.20C. The efficiency data for the nitrogen-ethylene dibromide system at 3-1/2 inch weir height and low gas velocity were not as reproducible as were the data at higher gas velocities or 1-1/2 inch weir. Of the five data points obtained in the low velocity range and at the 3-1/2 inch weir, three data points were low and two were high. This is the reason for two separate curves through the data in Figure 30. The reason for this non-reproducibility was not known. However, it was observed that the tray was operating in an unstable region at the low

-98velocities as indicated by the gas flowing through part of the caps at one time and through another part a few seconds later. There is a good possibility that the caps were leaking liquid during the time gas was not flowing through them. This could not be determined howevero This cycling was very apparent and it might be the cause for the scatter in the data since liquid leaking through the caps could have caused the inlet-gas sample concentrations to be in error. The efficiency data from the vaporization studies were converted to the number of mass transfer units by use of Equation (95) NOG =- 0n(l - EOG) (95) This equation applies in this case since there were no concentration differences along the length of the tray (pure liquid) andEMV = EOG, i.ed, plate efficiency equals point efficiency. In addition, the conditions on the tray were controlled so that the resistance to mass transfer was believed to be gas phase in which case E0G = EG so that Equation (95) may be written as follows: NG=- n(l - EG) (96) The data for the two systems are plotted in Figures 31 and 32. One point which should be remembered in regards to Equation(96) is that any error in EG is magnified in the resulting NGo This can be shown by the following relationship, dG= (1E) (97)

O 8.0 gpm, /L= 25 cp, I I WEIR 1.6 E 8.0 gpm, IL = 25 cp, 2 WEIR 8.0 gpm,/L =12 cp, 2" WEIR z 1.4 K 16.0 gpm, L/L 25 cp, 2" WEIR A 8.0 pm,L = 12 cp, 3 WEIR w 1.2 I: LL 0 0.8 wO 0.6 1.0 2.0 3.0 4.0 5.0 SUPERFICIAL VELOCITY, Us, ft/sec Figure 31. Number of Mass Transfer Units for Nitrogen - Cyclohexanol System. Data are Presented as a Function of Superficial Gas Velocity with Parameters of Weir Height, Liquid Rate, and Liquid Viscosity.

F WEIR 7.9 gpm, AVG. GAS TEMP. =27.7 WEIR 7.9 13 WEIR, *6 7.9 gpm, AVG. GAS TEMP. = 127J I1I 1.I | 7.9 gpm, AVG. GAS TEMP =115.50F, I2 WEIR A 15.9 gpm, AVG. GAS TEMPR =119.3 Fs 3 2 WEIR ~Z A A z A " 0.9 I - (1) 0 2 0.8 0 C: D 0.7 z 0.6 I I i 0.5 1.0 2.0 3.0 4.0 5.0 SUPERFICIAL VELOCITY, u.s, ft/sec Figure 32. Number of Mass Transfer Units for Nitrogen - Ethylene Dibromide System. Data are Presented as a Function of Superficial Gas Velocity with Parameters of Weir Height and Liquid Rate

As the point efficiency approaches 100 percent, any errors in the effi. ciency data are magnified many times in the NG data. An error analysis for the vaporization studies is presented in Appendix C. Mass Transfer Data - Absorption Studies The absorption studies.consisted of absorbing carbon dioxide in cyclohexanol at various gas rates, liquid rates, weir heights,,and liquid viscosities. The experimental conditions, are summarized as follows: I. 3-1/2 inch weir height. (a) Liquid rate, 4.95 gpnm; gas velocity, 0.9-3.5 ft/sec.,; liquid viscosity, 24.6 cp. (b) Liquid rate, 4.92 gpm; gas velocity, 0.9-3.5 ft/sec..; liquid viscosity, 5305 cp. (c) Liquid rate, 16.5 gpm; gas velocity, 0.9-3.5 ft/sec.; liquid viscosity, 52.5 cp. (d) Liquid rate, 26,.4 gpm; gas.velocity, 0.9-3,5 ft/sec.; liquid viscosity, 48.7 cp. II. 2 inch weir height. (a) Liquid rate, 4.94 gpm; gas velocity, 1.0-7.0 ft/sec.; liquid viscosity, 24.2 cp. (b) Liquid rate, 4.90 gpm; gas velocity, 2-7.2 ft/sec.; liquid viscosity, 97.2 cp. (c) Liquid rate, 16o5 gpm; gas velocity, 2.2-6.8 ft/sec.; liquid viscosity, 24.1 cp. (d) Liquid rate, 26.5 gpm; gas velocity, 2.2-4.9 ft/seca; liquid viscosity9 24.1 cp.

-102The liquid viscosity of the cyclohexanol was varied by changing the temperature of the liquid. The Murphree liquid efficiencies for the 3-1/2 inch weir height are plotted versus F-factor in Figure 33 and the data for the 2 inch weir height in Figure 34. These efficiencies are based on the liquid concentrations at point 5 on the tray shown in Figure 4 o A sample data sheet and calculation procedure are presented in Appendix Fo The detailed data are presented in Table -III-G. The data for -both weir heights are dependent on gas velocity or F-factor, liquid rate, and liquid viscosity. The increase in Murphree liquid efficiency with increasing gas velocity reflects the increase in gas holdup on the tray which provides more interfacial area for mass transfer and possibly an increase in the mass transfer coefficient. At a constant liquid viscosity an increase in.liquid rate causes the efficiency to decrease and probably is due to the greater flow rate and not so much to any change in the specific interfacial area, ft2/ft3, or in the mass transfer coefficient. Any rationalization at this point concerning the effect of liquid rate is complicated by the fact that the tQtal interfacial area increased as the liquid. rate was increased and the driving force also increased because of the lower level of concentrations in the liquid phase. The decrease in efficiency with increased liquid viscosity is no doubt caused by decreased diffusivity in the liquid phase and possibly by an effect of liquid viscosity on total interfacial area or specific interfacial area and the mass transfer coefficient

-10380 I -.70 60 z 0f I 50 U U. IL / 4030 20 *4.95GPM,/&L. 24.6 cp i 4.92 GPM,,.,, C 53.5 cP * 16.5 GPM,,LL.- 52A cp I0'10 —-_ / / T +26.4 GPM,/LL a 48.7 cp o..... I, I,...., 0 0.2 0.4 0.6 0.8 1.0 1.2 F- FACTOR Figure 33. Murphree Liquid Efficiencies for Carbon Dioxide - Cyclohexanol System - Variable Liquid Rate and Liquid Viscosity - 3-1/2 inch Weir, 4-inch Splash Baffle

70 TO 0 60 z w 0 w C.) lii 50 0~~~~~~~~~~~~~~~~ z 40 40 o 0 w 30 0 4.94 GPM, At~= 24.2 cp A 4.90 GPMqILL a 97.3 cp 0l 16.5 G PM,ILL a 24.2 cp D 26.5 GPN, /'L - 24.2 Cp 20 l0 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 -2.2 F- FACTOR Figure 34. Murphree Liquid Efficiencies for Carbon Dioxide-Cyclohexanol System Variable Liquid Rate and Liquid Viscosity-2-Incb Weir, 24.Inch Splash Baffle.

-105The conversion of the plate efficiency to point efficiency was not as simple in the absorption studies as in the vaporization studies. The expressions which have been used to relate plate efficiency and point efficiency were discussed in a previous section. The main thing to keep in mind is that these relationships are merely ways of accounting for the variable driving force due to the continuously increasing or decreasing concentration in the liquid as it flows across the tray. In the present study, concentrations at four different points on the tray were determined concurrently with each efficiency and these were used to determine an average liquid concentration on the tray, The average liquid concentration was used in the following relationship to calculate the point efficiency. EMV Xavg - yl/m Xavg- (98) EOG Xo - Yl/m xo - xt The value for EMV was calculated by use of the material balance relationship, Equation (62). Equation (98) was derived as follows: E -avg -y (99) mXo - Y1 where Yavg - the average concentration of the gas leaving the tray. Y1 = the concentration of the gas entering the tray. mxo = the concentration of the gas in equilibrium with the liquid leaving the tray~ BE~~~OG - r Y - yl~~ (100)

where y' = the concentration of the gas leaving at a point on the tray. mx' = the concentration of the- gas in equilibrium with the liquid at the same point on the trayo If Yavg is defined as follows, Yavg = f1 y'dW (101) Yavg o where w = fraction of distance across the tray, then Yavg = Jl [EOG mx' - EOG Yl + yl]dW (102) and Yavg = E0G mf0 xdW - EOG Y + l (103) if ESG, m, and Yl are assumed to be independent of the position on the tray. Representative concentration profiles for the studies at 3-1/2 inch weir height are presented in Figure 355 The concentration was found to be essentially linearly dependent upon the distance along the tray for all conditions studied. Since the distances between the four points on the tray were about the sae, it was thought that the integral in Equation (103) could be approximated by the arithmetic average of the four concentrations. That is, EG m x dW = EOG m [C2 + C3 + C4 + C (104) =G 0 (14PML =EOG m xav~ (105)

-107120 LIQUID RATE, 25.8gpm CL, 48.7cp I 10 WEIR HEIGHT,3 1/2 INCHES 18A, 185 Fo 100 _o oo11 E 90 I g 13A, 13!J 0 x w /15A RUN 1A 0. 70 0.0 3~~~~~~0~~0 0 F-FACTOR fRA<:TION OF: DISTANCE ACROSS TRAY 0 RUN 15A 0.287 * RUN 158 0.28? 0 0.2 0.4 0.6 0.8 1.0 Carbon Dioxide - Cyclohexanol. System

-10o8 where C2, C3, C4 and C5 are the concentrations at the four sample points shown in Figure 4 Equation (103) then becomes, Yavg- Y= E m Xavg - Yl) (1.06) If Equation (103) is substituted in Equation (99), the relationship between point efficiency and plate efficiency (Equation 98) is obtained The point efficiencies were.used to calculate the number of liquid transfer units by use of Equation (48) which is an equation combining the resistances in the gas and liquid phases and the over-all resistance to mass transfer0 Since the concentration of carbon dioxide in the gas phase was always above 50 percent and. in most cases in range of 70-90 percent, the gas phase resistance in Equation (-48) was neglected. The values of NL calculated in this manner are presented in Table IV-G and plotted versus F-factor in Figures 36 and 37~ In the derivation of Equation (95) which relates the point efficiency and the mass transfer coefficient, and. in the derivation of the relationship between point and plate efficiency, the assumption was made that the concentration of the liquid is independent of the vertical position in the froth at a fixed position along the tray. In order to establish the validity of this assumptions liquid samples were obtained by use of the sampling probes shown in Figure 4, These data plus the conecentrations on the tray floor are presented in Table VII and plotted versus F-factor in Figure 38.o The best lines through the concentrations on the tray floor are straight- and parallel, Sat indication that the rate of approach to complete mixing with increasing gas velocity is practicalely

NUMBER OF LIQUID-PHASE MASS TRANSFER UNITS, NL 0 0 * ~~~~ooe' ON t~~~~~~~~~~~~~~~~~iP'U~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ O IO tc D t..,-.... I-b X r (D Fj- pi 1-i~(DW 3 c, Hzm012 r FJ- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I II- rII 0o 0 o~~~~~~~~~~~~~~~~~~~~~~~ 0 Ca~ oI (D. 0 m ~ i'Tl 0H0 (D ) I-I* Hj-I H~o c o FH6CD Hi c+ 0 ti0 0 CO 0 CD1 -- _ 0

oO 1.5.J z Z U. Iz Lu Z Co 4 U) w - 4 o: H _ I1.0 ~o w " 0.55 0 0.2 0.4 0.6 0.8 1.0 1.2 IA 1.6 1.8 2.0 F -FACTOR Figure 37. Number of Liquid-Phase Mass Transfer Units for Carbon Dioxide - Cyclohexanol System - Variable Liquid Rate and Liquid Viscosity - 2-inch Weir, 21-inch Splash Baffle I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0........... F -FACTO~~~~~~~R ~~~~~~~~Fgr3.NubrofLui-h s as TraseUntfoCrbnDxie-ychxno Syste - Vaiabl iqud Rat andLiqui Visosiy-Ric er ~ic o~~~~~~~~~~~~pahBfl

..J _Iz UL. z 4 U),. C 4 Lu U) 4 0,. ~~~~~~oH x ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~ ~ ~ ~~ a ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~F-J _ I0 o 0 0 0.2 0.4 0.6 0.8 1.0 1.2 IA 1.6 1.8 2.0 F -FACTOR Figure 37. Number of Liquid-Phase Mass Transfer Units for Carbon Dioxide - Cyclohexanol System - Variable Liquid Rate and Liquid Viscosity - 2-inch Weir, 21-inch Splash Baffle

-1120.9 TRAY POSITION I 0 INLET SAMPLES 2 1 0.8 4 a SAMPLES FROM TRAY FLOOR 5 +J 1 FROTH l 08 A AMPLES + + 0.7 9.. + A ** I~''iA 0.6 z w z 0 Xoo 0.3 -Z IiJ ~ IC/I 00.2 z 0. 0 a. o 0.1 0 0.2 0.4 0.6 0.8 1.0 1.2 F- FACTORUs"G' Figure 38. Ratio of Point Concentrations and Equilibrium Concentration on Test Tray; 3-1/2 inch Weir Height; Liquid Rate, 4.95 gpm

-113nil. The concentrations in the froth are shown by the solid points in Figure 380 The sampling probes in the froth were positioned such that the concentration at position 7 should agree with the concentration at position 2, the concentration at position 8 should fall between the concentrations at positions 3 and 4, and the concentration at position 9 should agree with. the concentration at position 5 if there were no concentration gradients in the vertical direction.of the froth. If the three data points for position 9 at the high F-factorls are discounted as being in error, the concentrations at that position in the froth agree with the concentrations on the tray floor directly below. The data for position 8 are between the data for positions 3 and 4. Thus it is con-. cluded that no significant vertical concentration gradients were present in the froth in the area between the middle and the outlet of the tray, The concentrations at position 7 in the froth are higher than the data for the corresponding point on the tray floor. In fact, the concentrations are about the same as the data for position 8. The reason for these concentrations being about the same is believed to be caused by the recirculation of the liquid from the center of the tray to the inlet downcomer. This is partially due to the presence of the inactive area between the inlet downcomer and the first row of caps. The presence of a concentration gradient in the froth near the inlet downcomer does not discount the validity of the assumption of no gradient since the gas did not flow this area, In this regard, it would have been of interest to investigate the concentration in the froth above the first row of caps and not near the inlet downcomer. However, the results at points 8 and 9 in the froth verify the validity of the assumption for ~the major part of the tray and would seem to justify -the use of this assumption in the general case,

CORRE LAT I ON OF DATA Vaporization Studies Two different methods for correlating the vaporization data were considered. These were: (1) correlation by use of an equation of dimensionless groups and least-square fit to the data and (2) corre'lation by graphical analysis of the data using Equation (2-3a) as a mathematical model. NG = kda (Zf - Zc)/us (23a) The latter- method proved to be the more reliable method for this particular case~ The dependent variable in both correlations was the number of mass transfer-units. However, in the graphical method,~ the final form of the correlation can easily be converted to a correlation of the mass transfer coefficient. Both correlations are discussed in the following sections. In choosing the independent variables to include in the dimensional analysis, there was no question about including the following variables: gas velocity, us; gas density, PG; gas viscosity, Gt; gas diffusivity, DG; liquid density, PL; liquid viscosity, LL; surface ten — sion, aL; and a characteristic length, DS. However, there was some question whether the additional variables which are required to completely define- the system should be the variables, liquid rate and weir height, or gas and liquid holdup. It was finally decided that one additional variable, gas holdup, should be sufficient to complete the system of variables. The main reason for choosing gas holdup was because the hydraulic characteristics of this tray were believed to be affected by

-115the splash baffle and the liquid entering the tray. Had the variables liquid rate and weir height, been used, the exponents on these variables would not necessarily apply to larger columns and would therefore be misleading to anyone- not completely familiar with the characteristics of the tray used in the present stuady Of course, the use of gas holdup alone discounted the possibility of additional effects of liquid rate and liquid holdup. However, this omission is believed to be justified on the basis of the final fit of the data. In functional notation, the dependent variable, NG, may be expressed as a function of the independent variables as follows: NG = f[(Zf - Zc), Us, PG A1G, DG, PL, lL, aL, DS] (107) By use of dimensional analysis, these variables were combined to form the dimensionless groups in Equation (108). Zf-_Z a P b DUSPG c d e DaoLP f N = A c( 1G s ) (108L ) Ds PGDG 1G PG IG ((L)2 The values of the gas viscosity and diffusivity used in calculating the dimensionless groups in Equation (108) were predicted by use of the correlations of Bromley and Wilke(8) and Wilke and Lee(97), respectively. The liquid density and viscosity of cyclohexanol were determined experimentally and are reported in Tables I-E and II-E and Figures 1E and 2E in Appendix E. Since the ethylene dibromide was relatively pure, the liquid density and viscosity for this liquid were obtained from the literature. (47) Ashby(7) and Warzel(92) also studied the effect of gas phase resistance to mass transfer on the same tray as used in the present study. These data were included in the present correlations to broaden

-116the ranges of the several variables involved. In Table-VIII the gas and liquid properties are presented for the systems used in the author's studies plus those for the systems used by Ashby and Warzel. Ashby reported the physical properties of the systems used in his study and these are included in Table VIII. The properties for the air-water system used by Warzel were obtained from the literature.(67) The gas viscosity and diffusivity for all systems in Table VIII were calculated by the Bromley and Wilke(8) and Lee and Wilke(97) methods, respectively. In the actual process of correlating the vaporization data it was decided to start with Ashby's data and. determine the constants in Equation (108). Then Warzel's data were added to Ashby's data and another set of constants were determined. Finally, the author's data were added to Ashby's and.-Warzel's data and the correlation was performed for a third time. A summary of these correlations is presented in Table IX. It should be mentioned that these correlations were performed by use of an IBM 650 electronic computer and the regression analysis program prepared by J. G. Wendel was used in the computer to solve for.the coefficients in Equation (108) by the method of least squares. The reason for grouping the data and performing the correlation for each group was to determine whether the complete set of data could be correlated by Equation (108) and if not, to determine where it failed to correlate the data satisfactorily. The absolute average deviation between the experimental values and the calculated values of NG was not determined for each correlation. However, the standard deviation of the natural logarithm of NG was determined and. is presented fosr each correlation in -the last column o Table IX. The- smallest standard deeviation is

TABLE VIII SUMMARY OF GRAPHICAL ANALYSIS OF VAPORIZATION DATA PG G Nsc IL DG PL System lb/ft3 lb/ft-hr /PGDG cp t2/r lb/ft a + a = n C(1) Gn/2 pG/2 C"(2) 1. N2 - Cyclohexanol 0.0676 0.0464 2.98 1.84 0.231 57.8 -0.23 0.77 2.276 0.3544 0.260 1.670 2. N2 - Cyclohexanol 0.0708 0.045 3.02 24.14 0.210 57.9 -0.042 0.958 2.855 0.2834 0.268 2.70 3. N2 - Et Br2 0.0706 0.0456 1.90 1.45 0.341 134.0 -0.558 0.442 2.920 0.5568 0.266 1.395 4. Air- H20 0.06893 0.0472 0.56 0.764 1.21 61.7 -0.620 0.380 11.58 0.6015 0.262 5.04 5. He - iC4H9OH 0.01303 0.0504 2.20 2.38 1.74 49.1 -0.491 0.509 9.26 0.3311 0.114 3.188, 6. N2 - iC4H90H 0.066953 0.0451 1.54 2.310 0.438 49.1 -0.544 0.456 7.67 0.5398 0.272 5.865 7. Freon 12 - H20 0.2737 0.0351 0.24 0.598 0.522 61.9 -0.642 0.358 9.22 0.7930 0.525 6.081 8. He - H20 0.01048 0.0485 1.09 0.779 4.22 62.0 -0.666 0.334 17.09 0.4672 0.102 5.731 9. Air - NH3 - H20 0.07305 0.0434 0.613 0.851 1.04 61.7 -0.496 0.504 7.68 0.5171 0.270 4.010 O. He - MIBK 0.01636 0.0427 1.76 2.856 1.48 49.0 -0.595 0.405 10.93 0.4350 0.128 3.216 (1) NG = C usn-1 (Zf - Zc)072 (2) NG = C" (F-Factor)n-1 (Zf Zc5)072

TABLE IX SUMMARY OF VAPORIZATION DATA CORRELATION BY USE OF EQUATION (108) AND REGRESSION ANALYSIS (Equation 108) N A(GG G P0 Average Standard Average Absolute Absolute Maximum % Deviation of Constants in Equation A a b c d e f Deviation ~ Deviation Deviation Logarithm NG Ashby's data 5.502 0.713 (0.070) 0 -0.584 (0.046) -0.422 (0.028) 0.600 (0.024) 0.304 (0.013) 0.0580 Ashby's data 0.938 0.683 (0.073) -0.146 (0.100) -0.571 (0.046) -0.372 (0.044) 0.401 (0.139) 0.202 (0.071) 0.0580 H Ashby's plus Warzel's data 3070 0.588 (0.050) -0.834 (0.120) -0.480 (0.034) -0.125 (0.038) -0.600 (0.156) -0.321 (0.076) 0.1005 Ashby s, Warzel's, and author's data 4528 0.532 (0.040) -0.892 (0.020) -0.408 (0.029) -0.070 (0.023) -0.694 (0.042) -0.370 (0.020) 0.221 11.59 31.78 0.1170 Ashby's data 24.14 0.607 (0.072) -0.427 (0.020) -0.532 (0.047) -0.271 (0.028) 0.004 (0.008) 0 0.144 0.0612 Ashby's plus Warzel's data 10.86 0.556 (0.053) -0.341 (0.030) -0.441 (0.035) -0.236 (0.029) 0.056 (0.010) 0 0.1080 Ashby's, Warzel's, and author's data 6.279 0.548 (0.069) -o.696 (0.029) -0.514 (0.050) -0.129 (0.040) 0.056 (0.016) 0 0.304 19.64 77.80 0.2039 Numbers in parentheses are standard errors for the coefficients.

-119for the correlation of Ashby's data. The standard deviation for logarithm of NG is 0.0580 in this case. The standard deviation increased to,,0.1005 -when Ashby's.data plus Warzel's data were correlated and, to- 0Qll70. when the author's data were added to Ashby's and Warzel's data. The absolute average percent deviation for the latter case was 1L.59 with a maximum deviation of 31.78 percent indicating a fairly reasonable fit of the data. However, judging from the difference between the standard deviation of the logarithm of NG for this correlation and the correlation of Ashby's data alone, the correlation of Ashby's data is much better. The regression analysis program was planned so that the standard error for the constants in Equation (108) were determined.. In addition, in cases where the value of any one of the- constants was less than two times the standard error for that constant, the particular group involved was not included in the correlation. In the case of Ashby's data, the exponent on the Schmidt group was less than two times the standard error and the correlation was performed without this group. The correlation was also performed with the Schmidt number included by rejecting the test procedure -in the analysis program where the machine made the decision to drop this group. The standard deviation for the logarithm of NG in each of these cases was 0.0580. Thus it is concluded that the variable; gas diffusivity, DG; is not necessary to correlate Ashby's data satisfactorily. It should be noted, however, that the exponents on the remaining groups changed significantly when the Schmidt number was not included in Equation (108). Ashby varied the diffusivity in the gas phase by changing the liquid being vaporized and the gas used to vaporize the liquid. The physical properties of the gas phase for each of the systems used by Ashby are

-120included in Table VIII. By close examination of these data, a high order -of correlation between gas diffusivity and gas density can be seen. Therefore, in the correlation without the Schmidt number in Equation (108), the effects of the variations in gas physical properties are adequately correlated by gas density and viscosity and possibly the properties of the liquid~ This does not mean that gas diffusivity -per se is not an important variable. It does mean that with the type of equation used to correlate the data, gas diffusivity is not required to correlate the data satisfactorily. The variations of the constants in Equation (108) depending on the data used in the correlation are considerable.o The largest variations are in the exponents on the Schmidt number (-0.146 to -00892), the density ratio (0.o372 to -0.070), the viscosity ratio (0o401 to -0.694) and the surface tension group, (0o202 to -05370), The largest change in the constants in comparison with those for Ashby's data alone was found when Warzel's data were added to Ashby's data. The differences between the constants for correlation where the authorgs data were included and those for the correlation of Ashby's plus Warzel s data are not believed to be significant (see the standard errors in Table IX), Possible explanations for the variations of the constants in Equation (108) depending on the data used in the correlation are: (1) The data from the studies by War el and the author are in error considerable in relationship to Ashby's data. Of course, the reverse of this could be true. The statistician would make this statement by saying that one of the sets of data is biased. (2) The form of the equation used to correlate the data will not satisfactorily fit the data, (3) The variables

are not completely linear independent, or in other words, an orthogonal set of variables, therefore, the results presented in Table IX would be expected even for a slight bias in one or two sets of data. The partial correlation coefficients for the dimensionless groups in Equation (108) are presented in Table X. All three sets of data (Warzel's, Ashby's, and author's data) were used in determining these coefficients by the method outlined by Hildebrandt.(43) This method was a part of the regression analysis program used in the computer and the coefficients were a part of the data printed-out for each correlation. The square of any one of the numbers in Table X represents the approximate percentage of the variation in the dimensionless group at top of the particular column which can be accounted for by the group in the left hand column but in the same row. The reverse is also true. The highest order of correlation is between the surface tension group, D~aLp/(>2L)2 and the viscosity ratio, L/~G The partial correlation coefficient for these two groups is 0.933 indicating that about. 86 percent of the variation in either group can be accounted for by the other one. Because of this, the three groups of data were correlated by Equation (109) which does not include the surface tension group. The constants for the new correlations are also presented in Table IX. The NG A(fs) fG )b (DsPG) (pLd L)e (109) G Ds' PGDG PG G variations in the constants are not as great as in the cases where the surface tension group was included in the correlation equation. Also, the standard deviations for the logarithm of NG indicate that Ashby's and Warzel's data are correlated as well as in the cases where this group was

TABLEE X -PARTIAL CORRELATION COEFFICIENTS FOR THl: DIMENEL~SIONLESS GROUPS IN EQUATION10 D imesi on less GDUPGAlG PL'G D$2P/(I L)2 z Ze/DS iqQO -0.187 0.627 0.0117 0.029 0,027 CLG/PGDG pO.187 1.000 ~0.742 0.783...217 -0095 DSusPGhl~G 0,627 -oy742 1.000 -o.8i6 o00o4161 PJP -0. 117 7$5 — o': 16 1.;O. 115L 0.029 -0.247 0o048 -0.115 1.000 -0.933 nSarPL/(p~L) 0.027 -0.095 o*161 -0.08i -=.933 1.000

-123used, For the correlation where the author's data were included, the standard deviation for the logarithm of NG and the average absolute percent deviation for NG are both greater than in the case where the surface tension group was used in the correlation equation. The latter result indicates that the surface tension group does improve the correlation when the liquid properties, viscosity and density, are varied significantly. The greatest variation in the constants for the cases where the surface tension group was not included in the correlation equation is in the exponent on the Schmidt number. The range of variation is -0.341 to -0.696 and compares with the range of 0 to -0.892 in the case where the surface tension group was used in the correlation. These ranges of values for the exponent on the Schmidt number bracket the value of 0.5 which has been used by several investigators.(78,1,94) The value of two thirds has also been used however. (64,10) The results in Table IX indicate that Warzel's and Ashby's data are satisfactorily correlated by using the variable, gas holdup, in the correlation equation. Ashby's studies were conducted at 1-1/2 inch weir height while Warzel's studies were for 2 and 3-1/2 inch weir. Ashby(7) correlated his data by using in place of gas holdup, the height of liquid above the bottom of the slot opening which was determined by use of the height of liquid over the weir as defined by the Francis weir formula and an equation for slot opening as a function of gas rate and liquid density. The other groups in Ashby's correlation equation were similar to those used in Equation (108). Ashby used his correlation in an attempt to predict Warzelgs data for the absorption of ammonia in

-124water and found that the predicted values of NG were about 40 percent higher than the experimental values reported by Warzel, Ashby attributed this difference to'the correction for the liquid-phase' resistance which Warzel made for the ammonia absorption data. The liquid-phase resistance was estimated by use of Warzeli's data for the carbon dioxidewater system. Ashby reasoned that the small bubbles'in the froth which tend to be recirculated contribute more to the mass transfer in the' case of a low efficiency system (absorption of carbon dioxide in water) than in a high efficiency system (absorption of ammonia) where the gas is nearly saturated. He believed that this difference would cause the liquid-phase resistance for the carbon dioxide-water system when corrected for differences in liquid properties to be high in relation to the resistance which actually existed for the ammonia-water system,'The results of the present analysis do not support this reasoning since Ashbyvs and Warzel's data are satisfactorily correlated by either -Equation (108) or Equation (109). It is believed that the choice of the variable, L, by Ds Ashby was improper. The more direct variable seems to be gas holdup. However, AshbyVs correlation probably could have been improv hL eluding data over a wider range of the variable,. Correlations of the vaporization data by use of Equation (108) or Equation (109) are not satisfactory from the standpoint of being assured that the results can be extrapolated beyond the range in the variables studied. Although the constants in the equation were more reproducible when the surface tension group was not included, the correlation was not satisfactory for the authorgs data obtained for a system of high liquid density and systems oi' high liquid viscosities. For these reasons, it was decided that a graphical analysis of the data should be performed.

-125The graphical analysis of the data was approached by use of the relationship developed previously and used by Gerster(35) to correlate Ashby's data. This relationship was presented previously as Equation (23a). NG = kda - (23a) US Using the foregoing definition for klay Gerster correlated Ashby-s data as follows: kIa.= C uS023 (23b) where C = 18.19 DGO033 The data from the present study were used to calculate values of k a by use of Equation (23a) and the gas holdup and the superficial linear velocity for each run. The resulting data for the mass transfer coefficients are plotted in Figures 39 and 40. The value of the mass transfer coefficients for the nitrogen-cyclohexanol system are dependent on both the weir height and superficial gas velocity. The coefficients for 1-1/2 inch weir height are more than one and one-half times the values for the 2 inch weir and about two times the values for the 3 inch weir. The effect of weir height is not as great for the nitrogenethylene dibromide system in Figure 40. In this case, the values for the coefficients at the 3-1/2 inch weir height are about two-thirds of the values at 1-1/2 inch weir. The results in Figures 39 and 40 clearly indicate that the equation used by Gerster to correlate Ashby's data needs to be modified in order to correlate the data at the higher weir heights from the present study. Apparently, the successful correlation of Ashby's

-126100 90 80 70 60 50 40 30 0 z I-9 L in INCH WEIR U. w 10 -- 2 INCH WEIR 99U. I 7 ~$~3 INCH WEIR 4 3 NOTE = THE DATA FOR I 1/2 INCHES WEIR WERE ADJUSTED TO A VISCOSITY OF 12 CP. 2 -BY USE OF THE CORRELATION EQUATION, l -, I I I I I II', l! [,.I I I I' I 2 3 4 5 6 78 910 20 30 40 50" 00 SUPERFICIAL GAS VELOCITY, Us, ft/sec Figure 39. Mass Transfer Coefficients for Nitrogen-Cyclohexanol System at Three Different Weir Heights - Liquid Viscosity X 12cp.

-12720.0 * 10.0 e. D. 7.0 w, u5.0 - U00 0 M. 3-) 1/! INCH WEIR 2 m 2.0 1.0 1.0 2.0 3.0 5.0 7.0 10.0 SUPERFICIAL GAS VELOCITY, Us, ft/sec Figure 40. Mass Transfer Coefficients for NitrogenEthylene Dibromide System at 1-1/2 and 5-1/2 inch Weir Height.

-128data by Equation (23b) is due to the fact that these data were obtained at one weir height, 1-1/2 inch. In order to examine the- relationship between NG and gas holdup at a constant superficial gas velocity the data of Warzel(92) for the absorption of ammonia in water were used. These data cover an appreciable range of gas holdup at a constant gas velocity as shown in Figure 41. These data were used to determine the constants in the following equation by the method of least squares. NG = C usa(Zf - Zc)b (110) The constant, b, was found to be equal to 0.72 indicating that NG is not a linear function of gas holdup for these data. A similar relationship is suggested by the data from the present investigation in Figures 39 and 40. These data cover approximately the same range of gas holdup as covered by.Warzel's.data but the number of data points are too few to.establish a reliable relationship between NG and gas holdup. Therefore, it was decided to use the relationship established by use of Warzel's data and to determine whether this relationship would satisfactorily correct for the effect of weir height or gas.holdup on the data for the nitrogen-cyclohexanol and nitrogen-ethylene dibromide systems. The data of Ashby and Warzel plus the author's data were used to calculate the values of NG/(Zf - Zc)0'72 which are plotted versus F-factor in Figure 42. Since the gas density for each system presented in Figure 42 is essentially constant over the range of F-factor, the data of NG/(Zf - Zc)0'72 could have been plotted versus gas velocity with the

A 4.58 gpm, 2 WEI-R A 4.58 gpm, 3- WEIR * 32.04 gpm, 2" WEIRO 32.04 gpm, 3 WEIR 18.31 gpm, 2WEIR E 18.31 gpm, 3 WEIR2 ER * 9.16 gpm, 2 WEIR 0 9.16 gpm, WEIR zJ * z 3.0 I 0 3E 2. is * 1n 2.0 Z I 1 1 X 1 FigurebZ Inis U. 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 Zf -Zc, INCHES Figure 41, Number of Mass Transfer Units for Ammonia-Water System Data of Warzel(92)

10.0 N2- CYCLOHEXANOL, $584 /tr 5.0 2 O IoINCH WEIR, 8.0 GPM 2... * 2 INCH WEIR, 8.0 16.0 GPM | N2- CYCLOHEXANOL, j: 28.6 Ib/ft-hr O O 0 2 INCH WEIR, 8.0 GPM * 3INCH WEIR, 8.0 GPM 0 I0 N2- C2H4Br2 N 0 If1INCH WEIR, 7.9 GPM 2.0 NOTE: 3RINCH T WEIR,7.9 815.9 GPM 15 CHANGE OF SCALE AIR -WATER A a IJINCH WEIR, 8.OGPM A 2 INCH WEIR, 8.0 GPM 10 1 NH3-WATER (ABSORPTION 8 DESORPTION) 2 \ - 2 INCH WEIR 4.6 - 32.0 GPM t INCH WEIR, 4.6 -32.0 GPM 0 He - i C4H90OH I C3 IINCH WEIR, 8.0 GPM I- N2 - i C4H9OH N FREON-WATER Z I INCH WEIR,8.O0GPM * \O Ho-WATER | * $ NOTE: 1H I-INCH WEIR,8.OGPM CHANGE OF SCALE He - MIBK | -4 IINCH WEIR 8.0 GPM \.0 1I O 0.2 ~F-FACTOR,1.0 2 F-,FACTOR, Uvp-b Figure 42. The Effect of Liquid Properties Upon the Relationship Between NG0/(Z - Zc)0*72 and F-Factor

-131result being the same. The results in Figure 42 indicate that the use of the function NG/(Zf - Z,)0.72 does correct for the effect of gas holdup or weir height on the data from the present investigation. The deviations from the straight line through each set of data are appreciable in some cases but there is no definite trend with weir height. Another important observation concerning Figure 42 is the variation of the slopes of the best straight lines through the data. In fact, if the slopes of these lines are plotted versus kinematic liquid viscosity, the correlation in Figure 43 is obtained. The slope of the straight line through each set of data was determined by a least-squares fit of the data to the following equation, NG = C" (F-factor)a (111) (zf - Zc)0_72 The values of the constants, "a" and C", for each system are presented in Table VIII. The data in Figure 43 could be fitted to an equation of proper form but the form required would not be the most convenient to use in the final correlation equation. When the values of 1 + a for the different systems are plotted versus kinematic viscosity, the data can be represented by a straight line as shown in Figure 44. The plot of 1 + a versus liquid viscosity is also presented in Figure 44. The use of kinematic liquid viscosity appears to be the correlating variable since the data point for the high liquid-density system, nitrogen-ethylene dibromide, is better correlated by use of this variable with no obvious change in the correlation of the other data points. The best straight line through the data of 1 + a versus kinematic viscosity was determined by the method of

1.0 O 0.5- +V He -MIBK B N2 - CYCLOHEXANOL, 24.14 cp 0 N2-CYCLOHEXANOL, 11.84 cp N2-C2 H4 Br A AIR -WATER X He-iC4 H9 OH * N2-iC4 H9 OH 0.1 _ v FREON-WATER He -WATER + NH-WATER 0.04 - I I I I 0.5 1.0 5 10 20 30,LL/,L, cp/gm/cc Figure 43. Correlation of the Constant "a" in the Equation NG/(Zf - Zc)0'72 a/2 a C" PG us with Kinematic Liquid Viscosity

1.0 X He- i C4 H OH * N2- iC4 H90H y Hte —MIBK V FREON WATER O N2- CYCLOHEXANOL,24.14 cp A He-WATER 2 0 N2- CYCLOHEXANOL,I1.84 p + NHWATER' N - Cp H4 tB * CYCLOHEXANE - nHEPTANE A AIR-WATER U ACETONE-WATER /.L /PL' cp/gm/cc 4-1 o~~~~~0243 + x 0.3 0.3 1.0 10 20 30 40 /.L,C P Figure 44*. Correlation of "n" in the Equation KGa' = C"pG/2Uns=l+a/Zfcz)O-28 pGI/2 with Liquid Absolute Viscosity and Liquid Kinematic Viscosity

-134least squares. The correlation equation is, 09(L)0.238 (112) n = 1 + a = 0.393 (L)0.38 (112) where _L = kinematic viscosity, centistokes. PL Up-to this point the correlation of NG may be represented as follows, Nn-i n-1/2(f_ )072 (113) NG = C" Us PG (113) An equally valid correlation might be NG = C un-l (Zf - Zc )0.72 (114) The constants., C and C", for each system are presented in Table VIII. Equation (113) or Equation (114) may be converted as follows to a correlation of k~a by use of the definition, NG~kka US (23a) Therefore, k'a (Zf - Zc) = C" usn'l pGn-1/2 (Zf - Zc)0 72 and n n/2 k'a = C" U (116) PG Zf ) 8 By use of Equation (114), the relationship for the mass transfer coefficient is k'a C = - O 28 (117) (Zf - Zc) Equation (116) and (117) are identities and the constants may be related by the following equation if (F-factor)n-1 is considered to

-135be the correct way for combining gas velocity and density. C _ C" (118) G G However, if superficial gas velocity is considered as the correlating variable with the exponent, n-l, instead of F-factor, then the relationship between C and C" is not Equation (118) but some other relationship which can be determined only after C has been properly correlated. In the final analysis of the data, the use of F-factor and superficial gas velocity were examined. In order to complete the correlations, the variables required to correlate the constants C and C", were determined by trial and error. The variables which were believed to be necessary in these correlations were: (1) gas diffusivity, DG; (2) gas viscosity, LG; (3) gas density, PG; (4) liquid density, P.; and (5) liquid viscosity, LL. The range in the gas viscosity covered by the different systems is 0.0351 to 0.0504 lb/ft.hr. with the majority of the systems having a viscosity in the range of 0.0427 to o00485 lb/ft.hr. Therefore, in the first attempts to correlate the constants in Equations (116) and (117) gas viscosity was not included. The variable, C"/pG1/2, was first correlated as a function of the variables, DG, PL, and uL' The correlation is, _ _ ~ DGO.526 1L0.032 C" = 16.43 DGL (119) G — - pL0.83 where DG = gas diffusivity, ft2/hr. PG = gas density, lb/ft3. PL = liquid density, gm/cc. AL = liquid viscosity, cp.

-136The exponent on liquid viscosity in Equation (119) is not great and is probably not significant. For this reason, the correlation was repeated by using only gas diffusivity and liquid density. The equation for this correlation is as follows: C11"DG ~. D 526 C = 16.67 (120) The group, C" pL0.834/pG /2, is plotted versus gas diffusivity in Figure 45. The absolute average percent deviation between the values of C"/pG1/2 used in the correlation and the values predicted by Equation (120) is 8.28. The maximum deviations from the correlation line are for the two systems at high liquid viscosities (nitrogen-cyclohexanol). Equations (113) and (120) were used to predict the values of NG for the runs indicated in Table II-G. The absolute average percent deviation between the experimental values of NG and the predicted values using this correlation is 14.59 with a maximum deviation of 65.36. The corresponding values for the correlation of the data by Equation (108) are 11.59 and 31.78, respectively. An attempt was made to improve the correlation of the data by including gas viscosity in the correlation of C". This was done by arbitrarily including the gas viscosity and gas diffusivity in the correlating equation as follows, C" ~G1/2 G1 P1 ( LL P L) (121) G G The correlation obtained is presented as Equation (122), C" /2 = 3.14 Lo.023 (122)

100 EJ N2 -CYCLOHEXANOL, 24.14 cp ON2 -CYCLOHEXANOL 11.84 cp 0 N2 -C2N4 Br2 0.834 50 A. AIR -WATER CeP PL 0.526 = 16.674 DG X He -iCC He OH PG 2 C4H9H V FREON-WATER S IA He -WATER 0 j -co + NH3-WATER He - M'I BK ABSOLUTE AVERAGE PERCENT DEVIATION 8.28 1.0 0.1 0.5 V.3 5 10 - ft~'hr Figure 45. Correlation of "C' in the Equation u= n/2C1+a=nh with Physical Properties of Gas and Liquid

-138The deviations between the values of C" used in this correlation and the predicted values are greater than the deviations for the correlation by Equation (120). The absolute average percent deviation is 15.9 with a maximum deviation of 33.00 The corresponding values for Equation.(120) were 8.28 and 27.3, On this basis, the correlation of C" by Equation (120) was chosen as the best fit of the data. Therefore, the recommended correlation of NG using Equation (23a) is, NG = 523.9 26 G/2 (Ffactor)n liz f C )072 (123) pL~0.834 where n = 0.852 (8L)0.238 PL DG = gas diffusivity, ft2/hr. PG = gas density, lb/ft3. PL = liquid density, lb/ft3. TL = liquid viscosity, lb/ft.hro (Zf - zo)= gas holdup, ft. To determine whether the correlation of NG should be in terms of (F-factor)n-l or usn5l, the constants, C, for Equation (117) were correlated by use of the variables pG, DG, PL' and L' The result is as follows,.82 DG.405 o0,048 C = 10.582 DG0 6Pi 148 (124) o.816:0'.248, 4 CQ816 0.248 The function, -.RGO.04,, is plotted versus DG in Figure 46. The absolute average percent deviationbetween the values of C used in the correlation and the values predicted by Equation (124) is 13.746 The

I00 l,D peo -_~ C-~ ~10.582 d-J a0. 10 *- I0 ABSOLUTE AVERAGE PERCENT DEVIATION' 13.74 O13 Na-CYCLOHEXANOL, 24.14 ~p 0 Ne-CYCLOHEXANOL, II.84 cp Ng-CtH4Brt A AIR - WATER X He - iC4H9OH * Ng-iC4H9OH V FREON-WATER A He - WATER + NH3 -WATER V He- MIBK LO 0.1 1.0 10 D, ftt,/lt Figure 46. Corr-lation of the Constant "'C" in the Equation kGa' = C Usl+a=n/(Zf- ZC)0.28 with Physical Properties of Gas and Liquid

absolute average percent-..deviation between the experimental values of NG and the values predicted by use of Equations (114) and (124) is 21.08 with maximum deviation of ll5 percent. Therefore, the correlating variable, (F-factor)n-l, and not uTn-l, appears to be the one to use for the best fit of the data by the model represented by Equation (23a). Absorption Studies The correlation of the data for the absorption of carbon dioxide in cyclohexanol was accomplished with some degree of success by use of the following equation to define a mass transfer coefficient, NL = kL~ tL (39a) The final correlation consists of the mass transfer coefficient as a function of F-factor, liquid diffusivity, and.kinematic liquid viscosity. The major uncertainty in the correlation is correct value of the liquid diffusivity for carbon dioxide in cyclohexanol. The experimental values of diffusivity reported recently by Schoenborn(3) and values predicted by the semi-empirical equations of Chang and Wilke(l3) and Arnold(5) were used to indicate the different interpretations of the data which may be made depending on the diffusivities used in the correlation. Mass transfer coefficients were calculated by use of Equation (39a). The liquid contact time was determined by use of the following equation, (Zf AT - volume of caps) Zc/Zf tL (125= tQ L whereQ L = volumetric rate of liquid to the trayft3/sec. AT = area of the tray.

-141The calculated data are presented in Table IV-G in Appendix G and the mass transfer coefficients are plotted versus F-factor in Figure 470 The use of Equations (39a) and (125) to determine a mass transfer coefficient appears to correct for the effects of liquid. rate and weir height and certainly simplifies the correlation involved in accounting for these variables. The data for the liquid viscosity of 24.1 - 24.8 centipoises and variable liquid rate at 2 and 3-1/2 inch weir heights are satisfactorily correlated by a straight line with possible exceptions of five or six data points. The effect of liquid viscosity is very apparent by the three-fold. decrease in the mass transfer coefficient with a fourfold.increase in viscosity. The straight line through the sets of data at different liquid viscosities are parallel, however. Apparently, liquid viscosity does not affect the relationship between the coefficient and.F-factor as was found.for the gas-phase coefficients. Liquid diffusivities for the carbon-dioxide cyclohexanol system were first calculated by use of the correlation by Wilke and Chang(13) and used to correlate the mass transfer coefficients., The resulting correlation is presented in Figure 48. The exponent on diffusivity was taken to be 0.5, the value which was predicted by Higbie(42) and Danckwerts(22) for mass transfer through an interface which is continuously renewed. With the possible exception of the data at the high viscosity, 97.3 centipoises, the use of liquid diffusivity to the onehalf power corrects for the effect of viscosity. It was pointed out previously in the section entitled, "Experimental Procedure", that the viscosity of the liquid was varied by varying the liquid temperature. Therefore, these results indicate that the change in liquid viscosity

0 0.2 000~0 0.1 Ia,,~~~~~~~~~~~~~~~~l~ lb/ft-hr -j I-JR *4.95 GPM, 3-1/2-INCH WEIR L=59.4 0 4.94 GPM, 2-INCH WEIR,$L= 58.5 o3 16.5 GPM, 2-INCH W~EIRL =58.3, 26.5 GPM, 2-INCH WEIR,.,L =-58.4 O/~ &d 4.90 GPM, 2-INCH WEIR,/L =235.4 4.92 GPM,' 3-1/2-INCH WEIR, L.-29.3 i 16.5 GPM, 3-1/2-INCH WEIRpLL'12ZO + 26.4 GPM, 3-1/2-INCH WEIR, 1=,17,8 0. I ~ I! 0.1 0.5 1.0 5.0 10 F-FACTOR Figure 47. Mass Transfer Coefficients for Carbon Dioxide-Cyclohexanol System. Variables: Gas Rate, Liquid Rate, Weir Height, and Liquid Viscosity.

-143A 4.90 GPM, 2- INCH WEIR,jLc 235.4 Ib/ft-hr 0 4.94 GPM, 2^ INCH WEIR,/LL 58.5 Ib/ft-hr o 16.5 GPM, 2- INCH WEIR, L=58.3 Ib/ft-hr 0 26.5 GPM, 2-INCH WEIR, LL 58.4 b/ft-hr * 4.95 GPM, 3-1/2 INCH WEIR,g/L-59.4 Ib/ft-hr 4.92 GPM, 3-1/2 INCH WEIR, L 129.3 Ib/ft-hr 16.5 GPM, 3-1/2 INCH WEIR$, P1L. 127.0 Ib/ft-hr 26.4 GPM, 3-1/2 INCH WEIR,iL 117.8 Ib/ft-hr 0 I O O o, 0.5 0.1 I I 0.1 0.5 1.0 50O 10 F -FACTOR Figure 48. Correlation of Mass Transfer Coefficients for Carbon-DioxideCyclohexanol System

-144can be accounted for by a diffusivity correction. Also according to these results the effect of viscosity on the specific interfacial area or mechanism of transfer is not great. In order to compare the data for the carbon dioxides cyclohexanol system with the data for a system of low liquid viscosity, the data obtained by Warzel(92) were used. These data are presented in Table V-G. The Wilke and Chang correlation was also used to calculate the liquid diffusivity for carbon dioxide in water. The data are plotted as mass transfer coefficient divided by diffusivity to the 0.5 power versus Ffactor in Figure 49. Warzelis studies were conducted at a constant temperature at 770~F. Therefore, there should be no variations due to improper correction of liquid diffusivity for the data in Figure 49. The deviations in these data are considerablehowever, and by examination it can be seen that the deviations are a function of weir height. In fact, 1/2 at low values of F-factor, the values of kL /DL range from 0.21 to 0.36 x 102. The lower value is for the data at 3-1/2 inch weir height. This effect of weir height is similar to that found in the case of the gas-phase coefficients and could be caused by the same phenomenon. It is also possible that the. data were obtained in such a way to cause these variations with weir height. Maybe, the outlet liquid samples were not taken at the proper position on the tray. If the correlation in Figure 49 is compared with the one for the carbon dioxide-cyclohexanol data in Figure 48, it can be seen that the slope of the straight line through the data for the water system is smaller than that for cyclohexanol system. The difference is not great, however. Both sets of data (Warzel's and the authorYs data) were used to determine the constants in the following equation by the method of

ABSORPTION DESORPTION A 4.58 gpm, 3Z WEIR A 4.58 gpm, 2 WEIR 0 9.16 gpm, se us 0 9.16 gpm, a. -O- 18.31 gpm,.... U 18.31 gpm, a O 32.0 gpm, H * 32.0 gpm, a' V 4.58 gpm, 2 WEIR x 9.16 gpm m + 18.31 gpm " 032.0 gpm ii *b1 1.5CORRELATION LINE THROUGH DATA FOR CO -CYCLOHEXANOL SYSTEM'0 1. 0 X~~~~~~~~~~~~~~~~~~~~ 0.5/ 0 0.2~~~~~~~~~~~~~~~~~~~~~00. 0.1 0.5 1.0 1 0 F - FACTOR Figuare 49. Correlation of Mass Transfer Coefficients for Carbon Dioxide-.Water System.

least squares. NL kLa = 41.95 DL0478 (F-factor)0O575 (26) tL where DL = liquid diffusivity, ft2/hr. This equation is plotted in Figure 49 in order to compare the correlation line with the data for the carbon dioxide-water system. The absolute average percent deviation between the predicted values for both sets of data is 14.51 with a maximum deviation of 38.62. The absolute average percent deviation for the carbon dioxide-cyclohexanol system is 14.08 with a maximum deviation of 38.62. For the carbon dioxide-water system the corresponding values are 15.41 and 35.35, respectively. Therefore, the correlations for the individual sets of data are about the same. Although the foregoing correlation of the data is satisfactory, there is some question about its validity because of the uncertainty in the values of the liquid diffusivities as determined by the Wilke-Chang correlation. This is especially true for the diffusivities of carbon dioxide in cyclohexanol, a high viscosity liquid. The data used by Wilke and Chang in the correlation were for diffusion in dilute solutions where the viscosity of the solvent was not varied over a very large range. Gerster(38) used the data of Jordan, et, al.(49) and other data for systems of high liquid viscosities to demonstrate that the Wilke-Chang correlation is not accurate for systems of high viscosities. The Wilke-Chang equation is as follows, DL = 28.68 (xp. ) 1/2 (127) IB (VA)06

where VA = molal volume of the solute at normal boiling point, cc/gm-mole. MB = molecular weight of solvent. T = absolute temperature, K. LB = solvent viscosity, cp. XB = an association parameter to define the effective molecular weight of the solvent with respect to the diffusion process. DL = diffusion coefficient, ft2/hr. Equation (127) was used by Gerster(38) to predict the values of the diffusivity for oxygen in water-sucrose solutions and the results were compared with the data obtained by Jordan et al. (49) for the same system. These data are presented in Table XI. TABLE XI COMPARISON OF DIFFUSIVITIES FOR OXYGEN IN WATER-SUCROSE SOLUTIONS PREDICTED BY WIIKECHANG CORRELATION AND E IPEIMENTAL DATA BY JORDAN ET AL.(49)* Viscosity of DL, cm2/sec x 105 DL, cm2/sec x 105 % Sucrose Solution, cp Experimental Predicted 0 0.89 2.12 2.41 30 3.00 1.54 -0.848 54.5 25.0 0.67 0.124 62.5 67.o 0.43 0.0504 65.0 125.0 0.25 0.0278 * Table prepared by Gerster. ( 38)

The values of the diffusivities predicted by use of the Wilke-Chang correlation are about one-tenth of the experimental values, at the liquid viscosities of 67.0 and 125 0 centipoises, Incidently, the Jordan data were obtained by use of a polarigraphic technique. Gerster made a similar comparison for hydrocarbon systems with the results similar to those presented in Table XI. Several years ago, Arnold(5) presented a correlation for liquid diffusivities. Due to the more recent correlation by Wilke-Chang and also due to the fact that the equation contains some factors which are not readily available, the correlation by Arnold has not been used widely. The correlation equation was developed by extending a modified kinetic theory of gases to liquids and is similar to the Gilliland equation for gases. MA + MB 1/2'MA ~B'] DL = 0387 (128) where MA = molecular weight of the solute. MB = molecular weight of the solvent. =B - viscosity of the solvent, cp. VA = molal liquid volume of the solutes at the normal boiling point, cc/gm-mole. VB = molal liquid volume of the solvent at the normal boiling point, cc/gm-mole. AA = the abnormality factor for the solute. AB = the abnormality factor for the solvent.

-149Equation (128) is limited to a temperature of 200~C. For a temperature coefficient Arnold recommended the following equation. DL = DL0(1 + bt) (129) where b = O.O[20B/pL2/ 3 1/2 =tB = viscosity at 20~C, cp PL = density at 20~C, gm/cc The Wilke-Chang and Arnold correlations were used to predict the diffusivity of carbon dioxide in water at several temperatures. These data are compared with experimental data from several literature sources(46,l2,6165) in Figure 50. The values predicted by the WilkeChang correlation agree satisfactorily with the experimental data in the temperature range of 10 - 200C but at a temperature of 400C the predicted value is about 16 percent higher than the experimental values. The predicted values by Arnold are also in good agreement with the experimental values near 20~C. At higher temperatures the departure from the experimental values is significant however. In fact, the temperature coefficient suggested by Arnold is not satisfactory outside of the range of 15 - 2 5 ~C. Wilke and Chang reported association factors for four different solvents. These were: Water, x = 2.6; methyl alcohol, x = 1.9; ethyl alcohol, x = 1.5, and benzene, x = 1.0. Arnold reported values of AA and AB for several solvents and solutes. However, in both these correlations the principal problem is the determination of the association factors for systems besides those reported. In order to estimate the abnormality factor for cyclohexanolg an attempt was made to correlate

-15014.0 - 13.0C WILKE-CHANG CORRELATION DL - 28.68(XMJ12T L BVA'06 Xs= 2.6 / 12.0 TK /-,48, cp V, cc/gm- mol 11.0 1 0. 0 o - ox~~~~ A~~ / / <~~ARNOLD 9.0 / EQUATION AA= 1.O, A= 5.0 8 / EXPERIMENTAL DATA 0 STEFAN (46) /3 HUFNER(46) O CARLSON(12) 7.0- Aa RINGBOM (5.6) U PEACEMAN, VIVIAN (65) I 6.0 I t, I I.I 5.o J 10 15 20 25 30 35 40 45 50 55 TEMPERATURE, C Figure 50. Comparison of Experimental Diffusivities and Diffusivities Predicted by Wilke-Chang Correlation and Arnold Equation

-151the values reported by Arnold plus values determined by use of diffusivity data from the literature. The correlating parameter which was finally chosen was the parachor which is defined by Equation (130)o M cy1/4 IP]I = (130) PL The correlation is presented in Figure 51. The values of the parachors were obtained from the extensive list prepared by Quayle(69) and are presented in Table XIIo Two important observations in regards to Figure 51 are worth consideration: (1) For certain solvents the abnormality factor is equal to 1.0 at the solvent parachor equal to 200 and greater but increases with decreasing values for the parachor and (2) the abnormality factor for the hydrocarbon systems are less than 1.0 indicating a higher diffusivity than would be predicted by generalizing the data from substituted hydrocarbon or non-hydrocarbon systems. Gerster explained the latter phenomenon by the existence of immobility in the hydrocarbon systems at low temperatures which creates holes for the diffusing molecules to flow through at a greater rate than would be possible in the case where solvent molecules are in random motion. The value of the parachor for cyclohexanol is 248. The predicted value of the abnormality factor from Figure 51 is 1.0 but could be as high as 1.15. The latter value was used in Equation (128) to calculate the values of liquid diffusivities for carbon dioxide in cyclohexanol. It was assumed that the carbon dioxide does not associate with cyclohexanol which means that AA in Equation (128) is 1.0O. The calculated diffusivities are compared with the data reported by Schoenborn(3) in Figure 52. The variation of the calculated diffusivities with

-15 210.0 C WATER 0 ALCOHOLS 6.0 _~< ACETONE. ETHYL ETHER 5.0 t- &X AROMATIC HYDROCARBONS 4.0 0 + PARAFFIN HYDROCARBONS(99) A ACETIC ACID 3.0 vV7 ESTERS * CARBON DISULFIDE 0\ e CHLOROFORM, CARBON m2.0 09O GATETRACHLORIDE y BROMO BENZENE < HEPTANE (5.6) 0x 0 o — IDEAL SOLVENT 0 I_* BEHAVIOR, A= 1.0 I.O e -Frx-7- ---..... IF~ ~~~~ \ o~~r~CH4 IN C7H16 o I2 IN C4 H6 — CH4 IN C HJ2 OZ7 16 NOTE: CH4 N C24H5" ALL POINTS NOT OTHERWISE ETHYLENE DIBROIDE INDICATED ARE FOR DIFFUSION OF EITHER IODINE OR BROMOFORM ANISOL AND PHENETOL IN TE SOLVENTS (5.6) * OXYGEN IN GLYCEROL-WATER IN THE SOLVENTS 56)SOLUTIONS (49) OXYGEN IN SUCROSE-WATER 0.1 I SOLUTIONS (49) 30 100 200 SOLVENT PARACHOR [P] Figure 51. Correlation of the Solvent Abnormality Factor in Arnold's Equation with- the Parachor for Several Organic Compounds, Water, a-nd Aqueous Solutions * Arnold, J. H., J. Am. Chem. Soc. 52, 3937 (1930).

-153O ARNOLD'S EQUATION, AA=.O, AB=1.12, VA AND VS PREDICTED BY USE OF ADDITIVE VALUES OF LE BAS rO SCHOENBORN EXPERIMENTAL 0.5 POINTS 0.4 E.3I 0.:5 — // DATA RECOMMENDED BY SCHOENBORN (3 ) 0.2 /7 / 0.1 I I I l l 10 15 20 25 30 35 40 45 50 TEMPERATURE,~C Figure 52. Liquid Diffusivities for Carbon Dioxide-Cyclohexanol SystemComparison of Values Predicted by use of Arnold's Equation with Experimental Values of Schoenborn

TABLE XII SOLVENT ABNORMALITY FACTORS AND PARACHORS USED IN FIGURE 51 DiffusiOn of Iodine ( 8) Solvent AB [P Benzene 1,08 205.7 Toluene 099 6 a M-Xylene O. 97 284.5 Bromabenzene -0.97 257T-8 Chlor-oform l 0 183. 4 Carbon tetrachloride 0.94 219.9 Carbon disulfide o100 143 06 Heptane O.66 311.3 Ethyl Acetate 1oO6 216 Amyl Acetate 0.97 335.1 Anisole 1,14 265 6 Phenetl o1.08 303.7 Ethylene dibromide 1o 13 215.7 Acetylene tetrabromide 1.62 310G0 Acetic Acid 1.86 13102 Ethyl Alcohol 1.88 126.6 Methanol 2.325 88-.2 Water 5 38 54. 2 Diffusion of Borm(63) Ethyl Ether 0.89 211.9 Benzene 1.08 20507 Acetone 1.19 161o 7 Water 4 70 54,2 Methanol 191 93.2 Ethyl Alc0hol 2.04 126.6 Propyl Aldohol 1.36 1650 Amyl Alcohol 1,14 243.3

-155temperature does not agree with variation for the case of Schoenborn's data if the four values recommended by Schoenborn are considered. However, if the experimental value of 0.34 cm2/sec at 390C is used instead of the lower value of 0.30 cm2/sec at 41.60C, the slopes of the straight lines through both sets of data are parallel. The diffusivities calculated by use of Arnold's equation are 1.8 to 1.9 times the values reported by Schoenborn. If the Schoenborn data are accepted as being correct then the abnormality factor for the Arnold equation should have been about 2.08, a value which compares very favorably with the abnormality factors for methanol or ethanol. However, this value would not be predicted by the correlation of the abnormality factors in Figure 51. Therefore it is believed that the experimental values reported by Schoenborn should not be accepted without some question about their validity. In order to correlate the carbon dioxide-cyclohexanol and carbon dioxide-water data by use of the two sets of diffusivity data in Figure 52, the values of kLa at F-factor equals 1.0 in Figures 47 and 49 were used. These values are tabulated in Table XIII. The correlating equation in this case is as follows, kLa = (F-factor)'0 575 (131) where B =fX(DL) or ~(DL, AL' PL) Diffusivities from the two correlations in Figure 52 were used to calcu1/2 late the ratios, kL/DL / at F-factor = 1.0, which are presented in Table XIII also. If the diffusivity to the one-half power were the only variable required, the resulting ratios should be about the same regardless of the liquid. viscosity and density. However, the results in Table

TABLE XIII CORRELATION OF LIQUID-PHASE MASS TRANSFER COEFFICIENTS k sec-1 DL x 105, kLa kLa x -2 DL x 105 ka StT C L tlb L ft2 DLFco i0 —- x 10-2 /DLpL2 x102 DLO X 10-2 L 10 ~System T 0C 1b 3 ft-h F-Factor = 1.0 ft2/hr DLE( ft2/hr DL 2 IY V'Lft-hr' Carbon dioxidewater 25.0 62.2 2.16 0.0347 0.486 7.68 (3) 0.554 0.'103 Carbon dioxidecyclohexanol 38.0 58.4 58.4 1.0 0.125 1.96 (1) 0.282 0.282 1.10 (2) 0.378 0.378 H Carbon dioxidecyclohexanol 25.5 58.9 123 2.09 0.079 1.40 (1) 0.210 0.304 0.814 (2) 0.276 0.399 Carbon dioxidecyclohexanol 14.5 59.4 235 3.96 0.042 0.905 (1) 0.141 0.280 0.558 (2) 0.180 0.358 (1) Liquid diffusivity by use of Arnold equation, see Figure 52. (2) Liquid diffusivity - recommended by Schoenborn; experimentally determined; see Figure 52. (3) Liquid diffusivity by use of experimental data in Figure 50.

-157XIII indicate that the ratios, kLa/DL, decrease with increasing liquid viscosity. These results suggested the correlation in Figure 53 where kL /DL1/ 2 at F-factor equals 1.0 is plotted versus kinematic viscosity. Although the data point for the carbon dioxide-water system is over one and one-half log scales separated from the carbon dioxide-cyclohexanol system, a smooth curve through the latter set of data can be extrapQlated to the data point for the water system. Unfortunately, no data are available in the intermediate range of kinematic viscosity. However, the data obtained by Fairbrother are included in Figure 53 to indicate the small effect of kinematic viscosity on the mass transfer coefficient in the lowa viscosity range, 0.035 - 0.14. Fairbrother's data are for the desorption of carbon dioxide- from water and water-glycerol solutions at a constant gas velocity and liquid rate. The values of kLa were calculated by use of Equation (21). Fairbrother used a five-inch square column containing four plates with one 3-1/2 inch bubble cap on each plate. The diffusivities of carbon dioxide in glycerol-water solutions were determined by use of Jordan's data(49) for the diffusivity of oxygen in glycerol-water. These data were obtained at 250C, the temperature used by Fairbrother in his studies. A correction was made for the difference between the molal volumes of oxygen and carbon dioxide by use of Wilke-Chang correlation. The results in Figure 53 suggest that the effect of kinematic viscosity on the mass transfer coefficient, kLa, excluding the effect on liquid diffusivity, is a function of kinematic viscosity itself. Thus the correlation of f in Equation (131) is, =t D1/ (JO/PT (132)

70 50 40 FAIRBROTHER'S DATA (28) -O b - 0 DESORPTION OF CARBON DIOXIDE FROM WATER AND WATER-GLYCEROL SOLUTIONS AT A CONSTANT 0 0 GAS VELOCITY AND LIQUID RATE x o 0 v.J A OWT% GLYCEROL, tLL 0.896cp Od \ 12.2 WT% GLYCEROL, LL =1.21cp 1.0 N 33.2 WT % GLYCEROL,.L= 236 cp,, OF SCALE M 0.8 + 3.7 WT% GLYCEROL,.LL 3.74cp 0.7 - 0.6 - SLOPES =0.5 0 CARBON DIOXIDE-WATER (92) I; 4 35 P O CARBON DIOXIDE-CYCLOHEXANOL,ILL 24.1cp T CARBON DIOXIDE-CYCLOHEXANOL /LL50.8cp 0 CARBON DIOXIDE-CYCLOHEXANOL,/.L - 97. 1cp NOTE: SOLID POINTS ARE FOR DATA CALCULATED BY USE OF'oJ DIFFUSIVITIES REPORTED SCHOENBORN (3) Y~~~L~ ~OPEN POINTS FOR DIFFUSIVITIES BY ARNOLD EQUATION 0.1 I I 0.01 0.1 1.0 4.0 KINEMATIC VISCOSITY,, t/hr Figure 53. Correlation of the Ratios of the Liquid Mass TLransfer Coefficiernt and Diffusivities at a Constant F-Factor with Liquid Kinematil-. Viscosity - Data for Water anrd Cyclohexanol Sstems and I!0- erGlycerol Solutiorns

TABLE XIV FAIRBROTHER' S DATA FOR DESORPTION OF CARBON DIOXIDE FROM WATER AND WATER-7GhYCEROL MIXTURES kL8 -2 DL X 1OX, ft2/hr DL x105, ft2/hr 2kLa x 10 2 Wt. % Glycerol kLa Oxygen Carbon Dioxide D -2 JPLt (IL 4L~~~L 0 125 16.63 17.85 93.5 0.0349 17.5 12.2 99 15.20 14.21 83 o.o468 18.0 33.2 76 8.02 8.64 81.8 o.o866 24.1 43.7 57 6.04 6.50 70.7 0.134 25.8

-16owhere g' and aO are functions of LI/PLL The data for the cyclohexanol can be correlated by use of a = -0.5, however. This is shown in Figure 53 and in Table XIII where kLa/(DLpL)1/2 equals 0.29 or 0.38 depending on which diffusivity values are used. Etherington(27) reported a correlation of liquid-phase mass transfer coefficients which is similar to the one presented in Figure 53. The data obtained by Walters(90), Horton(44), and Fairbrother plus some data obtained by Etherington were used in the correlation. The viscosity range covered by these data is 0.4 to 13 centipoises. Hydraulic data were not obtained in conjunction with the mass transfer data so that the mass transfer coefficients were not determined by use of Equation (39a). Instead, Equation (21) was used and the height, h, was determined by summation of the following terms, hc = crest of liquid over top of weirs. hL = vertical distance from top of bubble cap slots to top of weir. hs/2 - vertical length of slots opento weir. Etherington also arbitrarily used the liquid diffusivity to the two-third power in the correlation which is represented by Equation (91). Other correction factors were also included. in the correlation to correct for such variables as liquid density, slot width, surface tension and concentration. An analysis similar to those made by Geddes(34) and Chu(17) was used to determine how most of the factors should be included in the correlating equation. The final correlation of the data indicates that the mass transfer coefficient corrected for diffusivity.to the twothirds power is proportional to the liquid viscosity to the 0017 power

-161in the range below liquid viscosity equal to 3.0 centipoiseso Above this viscosity the coefficient is proportional to the liquid viscosity to the 0.73 power. It is difficult to make a detailed comparison of Etherington's correlation and the correlation in Figure 53 because of the different exponent on liquid diffusivity and other factors in the correlating equation. The results as presented by Etherington indicate that the transition from a small effect of viscosity to a large effect occurs at about 3 centipoises instead of about 25 centipoises in Figure 53. Maybe this difference could be resolved by use of some recent diffusivities for the systems used by Etherington and by use of diffusivity to the one-half power. Mixing Studies Several relationships between point and plate efficiency have been developed mathematically but have not been tested thoroughly by use of experimental data. In the mathematical development of these relationships one common procedure has been used to relate the point and plate efficiency by use of a mixing parameter. This procedure is best described by Equations (99) through (103) and the solution to the differential equations which describes the liquid concentration as a function of position on the tray. The basic differential equations which include a mixing term (see Table II) have been solved by use of appropriate boundary conditions at the terminal positions on the tray. Then the resulting relations have been used in Equation (103) to obtain the relationship between point and plate efficiency. The final relationships include a mixing parameter in every case.

-162In the present studies, liquid concentrations at four different points on the tray were determined experimentally and were used to calculate an average liquid concentration on the tray. The average liquid concentration was then used in Equation (98) to calculate a point efficiency. This point efficiency was considered as an experimental value which could be compared with the values determined by use of some of the relationships presented in Table IIo The four relationships which were used in this comparison included the equations developed by Warzel(9), Crozier(l8), Robinson(72), and Lewis.(53) The point efficiencies for the carbon dioxide-cyclohexanol system determined by use of each of these relationships are presented in Table VI-G in Appendix G. In addition, the number of ideal mixing stages for the Gautreaux and O'Connell equation are included in Table VI'-G. The calculation of the number of ideal mixing stages was performed by use of -the following equation and a trial-and-error procedure. MV = 1 [(1 + OG.) (1]70) EOG XEOG n The point efficiencies determined by use of the Warzel and Crozier equations are within + 10 percent of the values calculated by use -of the average- liquid concentration on the tray. In the majority of the cases, the agreement is within + 5 percent. The values predicted by use of the following equation suggested by Robinson(72) do not agree as well with the experimentally determined point efficiencies, however, EMV el (133)..G. =-.1

-163where q is determined by the slope of the straight line through the concentration at various points on the tray when the data are plotted in x - x vs W, XO - XO the fraction of distance across the tray. The deviations are in some cases as high as 35 percent but the majority of the values are within + 10 percent of the experimental values. Robinson(72) has more recently recognized that Equation (133) is not valid for the general case since improper boundary conditions were used in the solution of the differential equation. Equation (133) was obtained by use of the boundary conditions; (1) W = 0 at W = o and (2) x = xo dx at W = 1. The relationship for the boundary conditions, d = at W = 1 and x = xo at W = 1, is presented in Table II and is believed to be correct solution to the differential equation. The excellent agreement between the experimental values of point efficiencies in Table VI-G and the values predicted by use of the equations developed by Warzel, Crozier, and Robinson is probably due partly to. the lack of sensitivity to changes in the mixing parameter at the values of XEOG for the cyclohexanol system. The sensitivity of Equation (67) to changes in the mixing parameter "Cs' at five different values of XEOG is shown in Table XV. The results in Table XV indicate that for a value of XEOG equal to 0.5 the change in the ratio, EMV/EoG, is 1.136 to 1.050 for a change in the mixing parameter of 2.0 to 5.0. The corresponding changes in the ratio, EMV/EoG, at XEOG equals 1.0 is 1.298 to 1.108 and at XEOG equals 2.0 is 1.728 to 1.229. The value of XEOG in the majority of the data for carbon dioxide-cyclohexanol system is in the range of 0.3 to 1.3. Therefore, the mixing parameter in Equation (67) could be in error considerably and still not cause an appreciable error in the point efficiency

TABLE XV SENSITIVITY OF EQUATION 67 TO CHANGES IN THE PARAMETER "C" AT VARIOUS VALUES OF XEOG EMV(1)'IC" EOG XEOG = 0.5.0 1.298 2.0 1.136 5.0 1.050 10.0 1.027 XEo = 1.0 1.0 1.728 2.0 1.298 5.0 1.108 10.0 1.050 XEOG 2.0 1.0 3.20 2.0 1.728 5.0 1.229 10.0 1.108 EEO~G = 5.0 1.0 29.4 2.0 4.480 5.0 1.728 10.0 1.293 (1) E eOG/C_ 1 "MVf/OG: XEOG/C predicted by use of this relationship. The relationships developed by Crozier(21) and Robinson(72) are of similar form to the one by Warzel(92) and the sensitivity to errors in the mixing parameter would be expected to be similar to that shown in Table XV. The results in Table'VI-G do however indicate that the liquid mixing can be satisfactorily accounted-for by use of the relationships developed by Warzel and Crozier and possibly the one developed by Robinson if XEOG is in the range below a value of 2.0. It is possible that at higher values of XEOG these relationships would not be valid. In any event,

-165the use of these equations to calculate point efficiency from plate efficiency would simplify the experimental procedure since only one concentration on the tray would be required instead of concentrations at several points. The studies by Robinson and Gerster(3) indicate that the relationship between point and plate efficiency is more complex than those developed by Warzel and Crozier, Robinson and Gerster used. an eddy diffusivity to describe the liquid mixing and develop the following equation, EMV 1 - eT - 1 __- = l ( _, _..(77) EOG (1+M)(2+M) (2T+M) where 2 M LS.=S2 DEZcBpM J~tL Equation (77) is more fundamental than the equations developed by Warzel and. Crozier and reveals how the complex empirical parameters in the latter equations can be expressed in terms of a more fundamental parameter, DE, the tray dimensions, S and B; liquid and gas rate, L and G; the slope of the equilibrium line, and point efficiency. In fact, Robinson(3) has developed the relationship between the Warzel mixing parameter,'iC", and the basic parameters in Equation (77). That is, C =MXEQG) (,EoG134) Therefore, in using Equation (67) to determine point efficiency from plate efficiency, the mixing parameter must be determined experimentally in conjunction with the experimental determination of plate efficiency

-166or a correlation of "C"' must be available for use in predicting the point efficiency by a trial-and-error procedure. Warzel determined the mixing parameters for the ammonia-water system used in his studies but made the mistake, of applying the same.: mixing parameters to the carbon dioxide-water- system. In.order to esti-e mate the point efficiencies for the carbon dioxide-water system used in the present studies, the mixing parameters for several systems were correlated by Equation (135). "" = A 4 VfPLB) ( (15) The correlation of "C" for the ammonia-water and carbon dioxide-cyclohexanol systems and.for four different syst~ems studied by Schoenborn, et al. (2) is presented in Table XVI. Although the standard deviation for this correlation is 1.33 which indicates that the scatter in the data is dioxide-cyclohexanol system agree very well with the experimental values as shown in Figure 54. This is another indication of the lack of sensitivity in Equation (67) due to changes in the mixing parameter in the low range of XEOGs. The values of the point refficiencies forw the carbon dioxidewaterx system predicted by use of the "C" correlationm in Table XVI and the values predicted by use of the parameters for the ammonia-water system as correlated by Warzel are compared in Figure 55. The results in Figure 55 indicate that the point efficiencies predicted by Warzel- are in errorbas much as 14 percent at the higher value s;:of XEoGo -The value of 14 percent error in point efficiency ist psmall however, ncompared teo the error in the

0.016 0 04.95 gpm,,L= 59.4 b/ft-hr, 3 WEIR E1 4.95 gpm, L 129.3 Ib/ft-r, 3 2 WEIR A / 0.014 - 16.2 gpm,p-127.0 Ib/ft-hr, 3 1WEIR A 25.8 gpm, I= 117.8 Ib/ft-hr, 3 2 WEIR / 0.012 -J 0.0104 z a. Y3 w 0.008o- ~ 4.95 gpm',L58.5 tb/fthr, 2"WEIR o 0 4.88 gpm, = L=235.4 lb/ft-hr, 2WEiR U 16.5 gpm, L.=58.3 lb/ft-hr, 2WEIR 26.4 gpm,pL=58.5 lb/ft-hr, 2WEIR 0.004 - 0.002 - I I 0.002 0.004 0.006 0.008 0.010 0.012 0.014 0.016 0.018 Eoe (PREDICTED) Figure 54. Comparison of Point Efficiencies Predicted by use of'C" Correlation and Warzel's Mixing Equation with Experimental Point Efficiencies

A 4.58 gpm, 3 WEIR 0 9.16 gpm, A 0.04 0- 18.31 gpm,. A 4.58 gpm, 2 WEIR * 9.16 gpm,, * 18.31 gpm, El * 32.0 gpm, " z 0 0.03 1 w U /: 0.02 o U N I 0.01 0.01 0.02 0.03 0.04 0.0 Eoo (IC" CORRELATION) Figure 55. Comparison of Point Efficiencies Predicted by use of Warzel's "C" Correlation and Point Efficiencies Predicted use of New "C" Correlation

8.0 A 4.58 gpm, 3- WEIR 0 9.16 gpm,. 7.0 8 18.31 gpm,, A 4.58 gpm, 2 WEIR * 9.16 gpm,. s6.0 U 18.31 gpm,, * 32.0 gpm,, d 0 4. -1 w 0 cr I h / O I \ 4. 3.0 N 43o 3 ~ ~ ~0.0'0 I I3 2~.00 1.0 / I I 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 C (PREDICTED) Figure 56. Comparison "C" Values Used by Warzel and Values Predicted by "C" Correlation C02 - H20 System

-170TABLE XVI CORRELATION OF MIXING PARAMETER "C" Data Used: NH3- H20....... }l Michigan Data C02-Cyclohexanol Methanol-Toluene Acetone-Water NCS Data(2) Cyclohexane-n-Heptane Mek-H20 Correlation:.....I II IV Zc -0.082 C 1 6 VfpLB BusPL 0.262 PL)-0.176 ( 0.162 Absolute Avg. Deviation = 0.969 Standard Deviation = 1.328 Standard Deviations for Exponents Group I "O.Q37 II 0.036 III 0.037 IV 0.026 values of the parameter "C" as shown in Figure 56. The percentage error in "C" as predicted by the correlation presented by Warzel is greater than 100 in some cases. The point efficiencies predicted by Warzel's correlation of "C" and by the correlation presented in Table XVI are listed in Table V-G in Appendix G. The efficiencies predicted by the latter correlation were used in the correlation of the liquid-phase mass transfer data.

-171The ratios of plate efficiency and point efficiency which were determined experimentally in the absorption studies can be used to determine the effect of different system properties and tray variables -on the basic parameters as defined by the mathematical developments by Robinson and Gerster. (3) This can be described by use of Equation (98) and (77). EMV x avg -x (98) OG o EMV x avg- x* 1 - e-(rl+M) el - 1 (2il+M -~ -, -21q -3LIM (77a) EoG xo - x ()+M) ) 4( M) Therefore, in functional notation, x.- - X = f M (M, EOG) (136) X0 _ x In Figure 57, the correlation between the logarithm of X a- and X X XEoG is presented. The data scatter appreciably but there is definite correlation with XEoG. In fact, the slope of the best straight line through the data as determined by "eye" is 0.12. In Figure 58 the ratio, X aY-x * / EOG 0.12 is plotted versus F-factor. The deviations from xO - X the straight line through the data is less than 10 percent in the majority of cases with no apparent trend.due to variables such as weir height, liquid rate, or liquid viscosity. In addition, the total effect of Ffactor on the ratio, IX ~av~g 5x*/(Qll2, over the range of 0.2 to 2.0 is tion (77) are related, e r results in Figure 58 indicate that the effect of this parameter is very small in the rage of XEOG covered by data.

0 4.95 gpm /L.= 59.4 Ib/ft-hr, 32 WEIR 0 4.94 gpm laL=58.5 Ib/ft-hr, 2 WEIR o 16.5 gpm,L=58.3 b/ft-hr, 2" WEIR (0 A 4.90 gpm,u.L =235.4 Ib/ft-hr, 2" WEIR * 4.92 gpm,. L =129.3 Ib/ft-hr, 352 WEIR U 16.5 gpm, /L =127.0 Ib/ft-hr, 32 WEIR () 0.3 - * 26.4 gpm,=LL =117.8 Ib/ft-hr, 3Z WEIR Q 0.X 0.O.0 0 C x x O 0.2 0I. m I' 0AL A A 0. 0.1 0.2 0.3 0.4 0.6 0.8 1.0 2.0 3.0 4.0 XE oG Figure 57. Correlation of the Average Liquid Concentration on the Tray with XEOG

* 4.95 gpm =LL 59.4 Ib/ft-hr, 3F WEIR A 4.90 gpm, = 235.4 lb/ft-hr, 2 WE o 4.94 p 58.5 lb/ft.hr, 2" WEIR =129.3 lb/ft-hr, 3 WEIR El 16.5 gpm "LL= 58.3 lb/ft-hr, 2" WEIR 16.5 gpm, =127.0 lb/ft-hr, 31WEIR L ~~~~~~~~~~~~~~~~~~~~~~~~~~~2 Z 26.5 gpm'IL= 58.4 lb/ft-hr, 2" WEIR A 26.4 gp1' 117.8 lb/ft-hr, 3 WEIR 1.5 c 0~~~~~~ w 0 ~~~~ ~~~~~PC x 0 ~ ~ ~ ~~ @ I A Q *~~~~~AElE 0 1.2~~~~~~~t XIX~~~~ ~~~~~ I I II.2 0.4; —-" —0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 F- FACTOR Figure 58. Variation of Xavg- X*/ (E.) 012 with F-Factor xo - e

-174This observation agrees very well with the predictions which may be made by examination of any of the point and plate efficiency relationships. Furthermore, the results presented herein indicate that the investigator should choose a system which would have a value of kXEOG above 2.0 in order to perform a critical analysis of the point and plate efficiency relationships. Some of the systems where the resistances are distributed between the gas and liquid phases would most likely fall in this range. In.Figure 59, the point efficiencies for the carbon dioxidecyclohexanol system are compared with the efficiencies predicted by use of the plug flow equation (Lewis equation) and the plate efficiency which would be the point efficiency if complete mixing were assumed. For the solid points in Figure 59, the ordinate should be considered the point efficiency predicted. by use of the plug flow equation For the open points, the ordinate is EOG = EMV. The point efficiencies for the 3-1/2 inch weir height are about centrally located between the values for plug flow and complete mixing except possibly at the higher liquid rates, In the case of data for the 2' inch weir height, the point efficiencies are more in agreement with the values predicted by use of the plug flow equation. The only major differences between the conditions on the tray with the 2 inch weir and the 3-1/2 -inch weir are the liquid viscosity and the liquid holdup. The data at comparable viscosities in Figure 59 indicate that the mixing on the tray was greater for the case of 3-1/2 inch weir than for the 2 inch weir. In other words, the mixing increases with increased liquid holdup, This is in agreement with the predictions which can be made by examination of Equation (77), That is, as the liquid holdup is increased, M decreases if IE does not change and in the limit as M

-INCH WEIR 2 - INCH WEIR 2 ~~~~~~~~~~0 O 4.95 gpm, ul 59.4 Ib/ft-hr 0 4.94 gpm, LLz 58.5 tb/ft-hr 4.92 gpm, L= 129.3 lb/ft-hr 0 16.5 gpm,/L= 58.3 Ib/ft-hr 0.020 0.020 0 16.5 gpm,UL=~ 127.0 Ib/ft-hr 26.5 gpmL.4 b/ft-hr 1~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~ 6.5 gpM, ~L.4 Ib/ft-hr 0 0 26.4 ipm,'ILL 117.8 lb/ft-hr 0 A 4.90 gpm,.k235.4Ib/ft-hr NOTE: FOR OPEN POINTS ORDINATE IS EoEv O Emv 0 0 FOR SOLID POINTS ORDINATE O O IS Eos =In (XEmv +1) 00.05 O.-Ol ais - O.o 0.010- 0.010an E o o w a ah o0 i A& OOO0.005 -0.005 0 A~~~~~~ 0~~~~~~~~~ c 0.005 0.010 0.015 0 0.005 0.010.1 Emv Euv a~~~~~~~~~~~~~~~~~0 I',_~ i~~~~~~~~~~~i~~~AO ~~~~~~~~~~0000.005 0-1 EMV ________h X EoXV3 ~ -XE06 XA,-EVG:)7.-X Figure 59. Comparison of Experimental Point Efficiencies With Point Efficiencies for Complete Mixing ana Plug Flow.

-176approaches zero, the ratio, -, approaches 1L0o Thus the mixing inEOG. creases as the liquid holdup increases until the liquid dif~usivity is afecated appreciably. Thereafter the liquid mixing would be expected to decrease with liquid holdup.

DISCUSSION OF RESULTS Vaporization Studies The most satisfactory correlation of the vaporization data which was obtained by the detailed analysis is represented by Equation (123). NG 52 DG526pG/2 (F-Fator)n -1(Z _ )0.72 (123);NGa- = 23.9 p~O.b34 where 0L 0.238 n = o.852 (P-) Using Equation (23a) as the mathematical model to define the relationship between NG, kGa, gas holdup, and superficial gas velocity, Equation (123) may be converted to an expression relatifg the -mass transfer coefficient and the independent variables. That is, NG = kG'a (23a) Therefore, kG a = 523.9 G26 (FFactor)n (L23a) pLO.834 (Zf-Zc)O28 Several important points in regards to this correlation are worthy of discussion in relation to previous and future investigations. These are: (1) the effect of gas holdup; (2) the exponent on gas diffusivity; (3) the use of F-factor as a correlating variable; and (4) the variable exponent on F-factor. It should be remembered that the coefficient, kG'a, is a product of two separate but not completely independent factors. The coefficient, kG, is related to the flow characteristics of the gas phase within the conlfines of the liquid. The flow characteristics are however dependent on -177

-178the shape and size of the vapor space in the liquid. The average shape and size of the vapor space determines the value of a, the interfacial area per unit volume of gas. Thus the two factors are related and it would be desirable for academic reasons at least to be able to study each of these factors separately. Several attempts have been rade to perfcrm this type of study. The problems involved in a study of this type are principally describing the geometry of the vapor bubbles or spaces in the liquid and the mechanism of mass transfer within the gas phase. In the present study data for gas-phase and iliquid-phase mass transfer were obtained and by use of both sets of data it is possible to make some qualitative observations as to the effect of such variables as gas rate, liquid and gas holdup, and liquid and gas physical properties upon each factor, kG and a. The relationship between the mass transfer coefficient and gas holdup on the tray as represented by Equation (123a) indicates a decrease in the coefficient with increased gas holdup. It should be mentioned that the variable, gas holdup, is not unique since liquid holdup plus gas hold. up, or variables such as liquid rate and:weir height could have been used to correct for gas and liquid holdup on the tray. The successful correlation of the data by use of gas holdup alone is attributed to the fact that the gas holdup and liquid holdup are related and that the gas holdup is independent of the means by which liquid holdup is varied, i.e., byvarying the liquid rate at a constant weir height or vice versa. The correlation might have been performed by arbitrarily assuming that the effect of gas holdup is in accordance with Equation (23a) Then any variations of the mass transfer coefficient with the holdup on the tray could have

-179been accounted for by use of the liquid holdup on the tray. The degrees of freedom in the system of variables which define the holdup on the tray can be enumerated by writing the relationship between gas holdup, froth height and clear liquid height. Zc Zf (1- 4)= Z Zc (137) If the liquid holdup and F-factor on the tray are fixed then the froth height and gas holdup are defined. By fixing the gas holdup and F-factor, the total holdup on the tray is fixed since the froth density, Zc/Zf, is almost totally dependent onF-factor. Therefore, the number of degrees of freedom is two and if any two of the variables, liquid holdup, total holdup (froth height), as holdup, and F-factor, are specified, the hydraulics of the system are defined. Gas holdup was used in present analysis for sake of simplicity since the.use of liquid holdup would have required a complete regression analysis. This would not have been simple due to the non-linearity of Equation (123a). The decrease in the mass transfer coefficient with increased gas holdup can be rationalized as a variatiQn in the specific interfacial area and the coefficient, kG. Crozier(21) suggested two different types of bubbling regimes as a possible explanation for the difference between the gas-phase data obtained by Ashby(7) at 1=1/2 inch weir height and similar data by Warzel(92) at 2 and 3-1/2 inch weir heights. Since these two sets of data were satisfactorily correlated by Equation (123) or (123a), there would have to be a continuous transition from one.bubbling regime to the other as the liquid holdup is varied in order to be able to correlate the data from the two separate regimes. Chu, et al. (19), have suggested

that the mass transfer coefficient as determined by Equation (23a) can be related to the individual coefficients in the separate froth zones on the tray. kG a tG = kGlal tG1 + kG a2 tG2 (138) where kGal = the coefficient for the bubble formation zone, tGl = the gas contact time in the bubble formation zone. ka2..= the coefficient for the froth and entrainment zones, -.!tG2. the. gas contact: time in,the froth and entrainment zones. By use of high-speed:,motion pictures, Chu, et al. (19) studied theieffect of the variables, slot subergence, slot area,.skirt clearance, liquid viscosity, and surface tension, on the average interfacial area. per unit volume of vapor,- al, and thee total.contact time,. tG.. The studies were conducted by -use of a single slot in two different bubble caps. The maJority of the tests were. conducted using an aluminum cap.with the following dimensions: 1/2 inch diameter, 1-1/2 inch high and 1/16- inch wall thickness. The slot size was varied between 1/8 by-5/8.. to 1/2 by 1-1i/4 inches. One set of runs was, made by use of a commercial 3 inch-diameter cast iron cap with a single 1-1/4) by 5/8 inch triangular slot. Slot submergenee was varied between,:O.5. and 4.0 inches; the liquid surface tension between 30 and 72,dynes/cm; and. liquid. viscosity between 0,4 and 1.0. centipoises. T: The effect of vapor: rate was studied over the range of 15 to 50 ft/sec, superficial slot velocity, while the liquid rate was varied over the range of 0.5:to 12.0 gal/min per inch of free plate width. It was

-181found that both the average interfacial area per unit volume of vapor, al, and the total contact time, tG, were primarily affected by the head of liquid on the slot. Below a slot submergence of approximately 2.5 inches of liquid, interfacial area was shown to decrease with increasing submergence and increasing slot area. Above 2.5 inches submergence, inter'acial area was found to be independent of submergence but a function of skirt clearance, liquid viscosity, and surface tension. No significant effect of air rate on the interfacial area was found. The effects of weir height and liquid rate were satisfactorily correlated by use of slot submergence which is directly related to these variables. Although the viscosity was varied over the narrow range of 0.4 to 1,0 centipoises, the interfacial area at slot submergences greater than 2*5 inches increased.approximately 27.5 percent as viscosity was increased over this range. The foregoing results apply to the effect of the several variables on the interfacial area and contact time in the bubble formation zone. The extent of this zone was found topvary with slot submergence and is the cause for the decrease in interfacial per unit volume of vapor with increase in submergence. At low slot submergence the growth period was fairly short and the bubbles reached the liquid surface while they were still relatively small. A bubble emerging from the slot continued to grow until its top reached the liquid. surface. At higher slot submergences the mode of formation changed and a second growth period.occurred.after a bubble broke away from the slot. In this case a thin channel connected the bubble and the slot, and the bubble continued to expand until the channel became.unstable and broke. Because of the channel, larger bubbles

-182were formed at the higher submergences with a subsequent reduction in area per unit volume of gas. Continued increase in slot submergence above a value of about 3 inches resulted in no change in the average time required for the channel to collapse and consequently bubble size then ceased to be a function of slot submergence. Calderbank(lO) studied the effect of submergence on the efficiencies for the adiabatic vaporization of water, ethyl alcohol, and isopropyl alcohol on a 12 inch-diameter plate equipped with a single 3 inch commercial steel bubble cap and a 6 inch sieve plate. Slot submergence was varied over the range 0.25 to 3.5 inches. Gas holdup data were not obtained during these studies. However, the values of the number of mass transfer units, NG, were found to be linear with submergence above a value of one inch. At lower values of submergence, the values of NG decreased very rapidly with decreased submergence. Calderbank extrapolated the straight line portion of the curves to zero submergence to obtain values of the NG for the bubble formation zone. The results by Chu, et al., indicate that this is an incorrect analysis since the extent of the bubble formation zone varies with submergence. Unfortunately, gas holdup data were not reported so that the data by Calderbank cannot be analyzed in a manner similar to that used in the present study. The variation in the mass transfer coefficient with weir height found in the present study corresponds with the variation in interfacial area reported by Chu, et al. For the nitrogen-cyclohexanol system, the mass transfer coefficients at the 1-1/2 inch weir height or slot submiergence of about one inch are more than oe and one-half times the values for the 2 inch weir (slot submergence about 1.5 inches) and about two times

-183the values for the 3-1/2 inch weir (slot submergence about 2.5 inches). For the nitrogen-ethylene dibromdie system, the mass transfer coefficients at the 1-1/2 inch weir height are about one and one-half times the values at the 3-1/2 inch weir height. Thus it appears that the effect of slot submergence varies with liquid. properties or possibly with combined effect of gas and liquid properties. It should be mentioned that the majority of data used in the correlation represented by Equations (123) and (123a) were obtained in the range of slot submergence, one to three inches, where Chu, et al., observed a continuous bubble growth until the surface of the liquid was reached. It is doubtful that this phenomenon occurred in systems used in the present study since froth height was in some cases greater than ten inches. It is very likely, however, that there was a bubble formation ze wherein the bubble was growing continuously until it became dislodged from the slot. In fact, visual observation under some conditions, particularly when high liquid viscosities were used, suggested a bubble formation zone. In this case, the mass transfer coefficients determined by use of Equation (23a) might be shown to be related to the coefficients for the separate zones on the tray as follows, kG'a tG = kGjal tG1 + kGa2 tG2 (138) If the gas contact time in each zone is defined as follows, tG1 = i/Us tG2 = f/Us where l1 = gas holdup in bubble formation zone, =f = gas holdup in froth zone. f = total gas holdup.

Then substituting in Equation (138) = kGla. (f) + kG2a2 Pf (139) kG'a = kGLal (1 f) + kG~a2 13 (140) Therefore, at any fixed gas rate, the value of kGta depends on the ratio of the gas holdup in froth zone and th ttotal holdup. As the gas holdup in the froth zone becomes a large fraction of the total gas holdup, the value of kG'a approaches the value of kG~a2 9 the mass transfer coefficient for -the froth zone. If the froth zone is small compared to the bubble formation zone, the mass transfer coefficient approaches kGlalt The data from the present study indicate that kG al is greater than kG;a2 since the value kG a decreased as total gas holdup increased~ In the case of the nitrogen-cyclohexanol system, kGal1 is greater than two times the value of k Ga2o Equation (140) applies to the case where both the bubble formation and froth zones exist on the tray0 At low submergence of the submergence of the slots, the coefficient, kGlal. accounts for a large percentage of the total coefficient, kG'a. At very low slot submergence, it is very probable that the bubbles do not develop fully before they break through the frotho Thus the maximum bubble size which Spells and Bakowski(83,84) and Chu et alo (19) observed would not be obtained0 In any event, the total coefficient, kGca, would probably be better correlated by use of a relationship such as Equation (140)o If the results reported by Chu et alo are accepted as being applicable for the case of multiple slots, then the interfacial area in kGoal would be independent o: gas rateo The coefficient, kG, in

-185kG'al would be expected to be a function of gas rate however. The maximum bubble size or the interfacial area in the froth zone would be a function of gas rate if the results of Spell and Bakowski are accepted as being applicable. Therefore, the total coefficient, kG'a could be represented by the sure of two terms but each term would be a different function of gas rate. That is for any system, kG'a = A (F-factor)a(1- f) + B(F-faetor)b pf (141) One of the uncertainties in an analysis of this type is the extent of the bubble formation zone. If this zone is not fully developed in all cases of submergence then kG'al becomes a function of submergence and the present analysis is not valid. The dependence of the mass transfer in the gas phase upon the physical properties of the gas has been a major cause of investigation.since the advent of the Whitman film theory. According to this theory the mass transfer coefficient should be proportional toithe first power of -diffusivity. The effects of physical properties such as viscosity and density are not readily apparent unless it is rationalized that these variables affect the thickness of the film across which the transfer is occurring. Chilton and Colburn(l5) used an analogy to heat transfer to derive the following relationship between the mass transfer coefficient and the friction factor, f, which is calculated using the skin friction. JD - kGPBM (JG )2/3 = 1/2 f (142) GM PGDG Sherwood and Pigford(78) suggest the use of the Schmidt number to the minus one-half power by analogy with results from packed and wetted-wass column s.

-186The mass transfer coefficients for wetted-wall columns have been correlated by the following equation, kG -D PBM 0.023 ( )083 (8 PG )0 44 (143) DG - DG P PG' PGD)G For dilute mixture of solute gas at 1 atmosphere total pressure, Equation (143)' reduces to OGo56 0 83056" o ~I. 1 ($144) kG = 0.023 G. D0ll7 (144) C1G DO` PG or 0.56 7(Ffact0.r)Oo83 or.kG= 0.0253 < (F-factor). (145) LGSo39PGO.02 DO17 The important point to emphasize in regards to Equation (145) is that when F-factor is used as a correlating variable, kG is proportional to the gas diffusivity to the 0~56 power and not necessarily the Schmidt number to the minus one-half power. It is also interesting to compare Equation (145) with Equation (123a)o The dependence of kG and kG~a on F-factor diffe.r which might be expected, but the dependence on gas dif~ fusivity is about the same in both cases, The effect of the gas physical properties on the mass transfe: in packed towers has been studied by several investigators. However, the results have not been consistent. Mehta and Parekh(12) vaporized water, methanol, benzene, and toluene and found that H. To.Uo was proportional to the gas diffusivity to the 0o.17 power. Smosky and Dodge(8,) used water, methanol, benzene and.ethyl butyrate to study vaporization of these liquids and correlated. their data by use of gas diffusivity to the 0o15 power. Johnstone and Singh(48) found that rates of absorption of sulfur dioxide in sodium hydroxide solution; absorption of ammonia in aqueous acetic acid;

and evaporation of water into air from wood grids varied according to the Schmidt number to the minus two-thirds power. The data of Simkiin(80) and Chrisney(l6) on the vaporization of solids and liquids from packed beds in the absence of liquid flow indicate that the H. T. U. is proportional to gas diffusivity to the -0.36 power. More recently Lynch and Wilke(54) vaporized water into air, helium, and Freon-12 in a 12 inchdiameter packed tower to study the influence of gas properties on the rate of mass transfer. When these data were correlated by use of Reynolds number, DG/1G, H. T. U. was found to be proportional to Schmidt number to the 0.90 power. However, when the same data were correlated by use of Ffactor, the values of H. T. U. at a constant liquid rate was found to be proportional to Schmidt number to the 0.47 power. Lynch and Wilke also correlated psychrometric data from several sources(54) and found that the data could be correlated suitably by,-use of Schmidt number to the 0.5 power. On the basis of the similarity between the two types of data, psychrometric and packed column, Lynch and Wilke preferred the correlation of the packed column by use of F-factor and Schmidt number to the 0.47 power. It was also emphasized by these investigators that the transfer of mass and momerntum.might be expected to be functions of the same properties of the flow field. Furthermore, they pointed out that the drag coefficient at high values of Reynolds number is independent of Reynolds number and the pressure drop is dependent only.upon the inertia of the gas stream, pGu2. Since it was postulated that the rate of mass transfer will depend on the same flow properties as for the case of pressure drop, Lynch and Wilke concluded that the mass transfer coefficient would be a function of the gas inertia and not Reynolds number at high flow rates.

-188The results reported by Lynch and Wilke(54) are very similar to the results from the analysis of the vaporization data inthe present study. When a Reynolds number was used as one of the dimensionless groups to correlate the values of NG? the exponent on the Schmidt number was found to be -0.89 when all the data were used in the correlation~ However, when F-factor was used as a correlating variable, the number of mass transfer units was found to be proportional to the gas diffusivity to the 0.53 power. The number of mass transfer units might have been correlated by use of Schmidt number to the minus one-half power. However, the fundamental variable, kG@a, would definitely not be correlated by use of Schmidt number since the gas density is eliminated when Equation (123) is converted to a correlation of the mass transfer coefficient. It should be mentioned that the variation of gas viscosity in the data obtained by Lynch and Wilke was not great and the correlation would probably not be changed significantly by either including or excluding gas viscosity. The important point is that the kkGja is not a function of Schmidt number although the data in the for of Ho T. U. for packed towers and NG for bubble cap plates may be shown to be a function of Schmidt number. It has not been widely recognized that the penetration theory proposed by 42igbie(Q2) or Danckwerts(22) may be used to relate the gasphase mass transfer coefficient and gas diffusivityo However, the results Of the present study suggest that this theory is applicable for the gas phase. The penetration theory is based on a periodic renewal of fluid elements at the int-erface and it is not difficult to visualize such a process in view of the rb enc a ole in tb e gas phase in the bubble formation zone. In the froth zone, this renewal coul be caused by the

-189circulation of the gas in the bubbles due to drag at the interface. Rybczynski(74) and Hadamard(40) applied Stoke's method to "creeping" motion of a fluid sphere of zero surface tension to derive a relationship between the rate of circulation inside a fluid drop, the velocity of the drop, and the ratio of the viscosities of the inner and outer phases. 32 1 + IjO/I Hughes and Gilliland(45),used this relation to show that for a bubble the circulation has a great effect on the drag. Or vice versa, the drag on the bubble has a large effect on the circulation inside the bubble. Hughes and Gilliland concluded that bubbles of gas in liquids at high Reynolds number are always circulating. Garner and Hammerton(33) used ammonium chloride to observe the circulation inside bubbles. They concluded that bubbles greater than 0.03 cm in diameter in water tend to have internal circulation and that the energy caused by skin friction must be greater than the surface energy before circulation can persist. The skin friction would be expected to be a function of liquid viscosity. Thus the rate of circulation would be expected to vary with liquid viscosity. The use of F-factor to correlate the gas-phase mass transfer coefficient is believed to be justified on the basis of the discussion presented by Lynch and Wilke. However, in the case of the gas flowing through the liquid on the tray, the rate of momentum transfer is not related to the total pressure drop across the tray but the pressure drop due to drag between the gas and liquid. An energy balance on the gas between a point below and above the tray may be used to indicate the different causes for dissipation of energy.

-19ou2 2 -ZF-W=+ —-+ f VdP (147) 2g 1 If the work term ls neglected and the vapor density is assumed to be constant, then 2 2 P P u2 - + (Z2 Z)P = +ZF (148) PG 2g PG If the kinetic energy above and below the tray are not significantly different, then &PT = ZcPL + PG ZF (149) where -APT - total pressure drop. Zc = slot submergence. )F = frictional energy lOsses, ft.-lb/lb. PL = liquid density, lb/ft3. PG= gas density, lb/ft5. The term, ZF, is the total energy loss due to the gas flowing through the caps and the liquid. The energy loss due to the gas flowing through the liquid is the term which is related to the momentum transfer. Since this energy loss is probably small, an accurate measurement of the total pres — sure drop, the pressure drop across the caps, and slot submergence would be required before the —energy loss in the liquid.could be determined accurately. It would be of interest to see a study of this type performed in the future since the rate of mass transfer and pressure drop in the liquid might be related by use of the present theories of mass and momentum transfer. The gas-phase and liquid-phase data from the present study may be used to amplify the foregoing discussions in regard to the effect of liquid viscosity on the drag between gas and liquid and the relationship

-191between pressure drop and mass transfer. In the case of the mass transfer data for the gas phase the coefficients were correlated by Equation (123a) wherein the relationship between the coefficient and F-factor was found to be a function of the liquid viscosity. However, the liquid-phase mass transfer coefficients were correlated by use of Equation (131) wherein the relationship between the coefficient and F-factor was found to be independent of liquid viscosity. Since the relationship between the coefficient and F-factor was found to be a function of liquid viscosity in the case of the gas-phase data and not in the case of the liquid phase data, this is good reason to believe that viscosity affects the drag between the liquid and gas and consequently the mass transfer in the gas phase. If the effect of liquid viscosity were an effect on interfacial area then liquid viscosity should appear in the correlations of the gas and liquid phase data in identical form. Absorption Studies The correct correlation of the liquid-phase mass transfer coefficients is somewhat uncertain due to the uncertainty in the diffusivities for carbon dioxide in cyclohexanol. However, in view of the failure of the Wilke and Chang correlation to predict diffusivity data in systems of high liquid viscosities, it is believed that the correlation obtained by use of the diffusivities predicted by the Wilke and Chang relationship is not a correct representation of the data. Although Schoenborn et al.(3) determined experimental diffusivities for carbon dioxide, it was pointed out in a previous section that there is good reason to doubt the validity of these data. Arnold's equation was used to predict diffusivities for the carbon dioxide-cyclohexanol system which were 1.8 to 1.9 times the

-192values reported. by'ShchoE.:Iorri,:.', form. Qo of te corr elabion equation is as follows when either the Schoenborn d.a'ta or ithe diata predicted'by Arnold's equation are used. kLa, D L/2 (~sJipL)C (F-factor)0.575 (Dia) where f and l are functions of t/P The absolute value of pI depends upon which diffusivity data are used as shown in Figure 53. However, since the same form for the correlating equation is obtained by use of diffusivilty data from the two separate sources it is believed that Equation (l31a) is the correct correlating equation if it is assumed that -the penetration theory proposed by Highbie (42) and Danckwerts (22) is applicable. The important points to consider in regards t;o Equation (131a) is tthe variation of a with the kinematic liquid viscosity, the dependence of the coefficient on F-factor, and the relationship between the coefficient and liquid diffusivity. The same problem is confronted. with the liquid-phase mass transfer coefficient as in the case of the gas-phase coefficient. That is, the coefficient is a product of two separate but not unrelated terms. However, for the case of the liqu.id-phase coefficient, it is possible to predict the combined effect of the liquid properties and the gas rate if it is assumed that the gas flows through the liquid in a form approximating a bubble and. that Higbie s(42) equation describing the coefficient is applicable. This could not be done for the case of the gas-phase coefficient beca~use of the inabilit;y to des~cribe k@ in t~erms of the fluid dynamics and physical properties of the system HiLghbie's equation relates the liquzdwphase diff~slivity a:rd the bubble characteristics as follows:

-193kL = Lt (15a) where DL = the liquid-phase diffusivity. tb = bubble contact time, db/Vb. db = bubble diameter. Vb = bubble velocity. Van Krevelen and Hoftijzer(88) used data from several sources to perform a study of the characteristics of bubbles formed at orifices. Bubble diameter and velocity were correlated with liquid properties by use of the following dimensionless relationship. gdb P= M(Vb db PL) (150) Vb PL! AL In the streamline region of flow the above relationship was found to be as follows: gdb P (Vb db PL)-l (15Oa) Vb2 PL ML Therefore the expression for bubble contact time becomes, tb db 18 d L (151) Vb g db PL if ~p is approximately equal to pL. The maximum bubble diameter may be related to the volumetric gas rate and bubble frequency. QG = Nsnb (152) where nb = bubble frequencyo Ns = number of bubble sources.

-194and db= (6/NsnbL (152a) TNsnb or 1/db = (3Nsnbrl/. (152b) The interfacial area per unit volume of gas is equal to the reciprocal of the bubble diameter, i.e., a = 1/db. If Equations (151) and (152a) are substituted in Equation (15a), an expression is obtained which relates the mass transfer coefficient, liquid diffusivity, liquid,kinematic viscosity, and volumetric gas rate. That is, (tNsnb)16 (1/2 -1/(152 kLa = 1.51 () 1/2DL (153) 64G PL Two points in regards to Equation (153) are significant: (1) The mass transfer coefficient is proportional to the liquid kinematic viscosity to the minus one-half power in the range of high liquid viscosities and (2) The mass transfer coefficient based on interfacial area per unit volume of gas is not highly dependent on gas rate. The above observations are valid only when the bubble frequency is independent of gas rate and liquid properties. Calderbank used a special technique to study the frequency of bubbles found at slots and orifices. Over.the range of flow rates normally used in industrial practice, it was found that bubble frequency was constant at 20 per second and independent of gas flow rate, slot dimensions, spacing between slots, physical properties of gas and liquid and slot submergence. If the value of 20 per sec for the bubble frequency is substituted in Equation (153), the expression for the liquid-phase mass transfer coefficient in the high range of liquid viscosities becomes, kLa = 2.23 (N5)l/6 DLl/2 (LL)-l/2 (153a) QG:L

-195In the low range of liquid viscosities, van Krevelen and Hoftijzer found that the relationship between bubble diameter and velocity became, g.b..... 2 (15Gb) Vb2 PL Thus the expression for the bubble contact time is, tb = db 2Vb (154) Vb g and the expression for the mass transfer coefficient becomes, T a = 2(g 114 3/4 1/2 kLa = 2() db DL (155) If Equation (152a) is substituted in Equation (155) the dependence of the mass transfer coefficient on the gas rate is found to be kLa = 2.23D ()/4 1/255a) Therefore in the low range of liquid viscosity, the coefficient is dependent on liquid diffusivity and gas rate but independent of liquid kinematic viscosity. By comparison of Equations (153) and (155a), it might be predicted that the dependence of the coefficient upon the gas rate in the low range of liquid viscosity differs from the dependence in the high range of liquid viscosity. The interfacial area in these two equations is the specific area based on the gas holdup while the specific area in the correlation represented by Equation (131a) is based on the liquid holdup. The latter equation may be converted to a basis of area per volume of gas by multiplying by the ratio of liquid and gas holdup as follows: kLa = kLa Z Zc (156) Zf c Z

-196In the preliminary examination of the data.for the carbon dioxidezc cyclohexanol system it was found that the ratio, c, could be correlated by use of F-factor to the minus 0.75 power. For the case of the carbon dioxide-water system, the ratio was found to vary with F-factor to the minus 0.6 power. The ratio in both cases however was found to be a function of weir height and liquid rate. As shown in Figure 49 the mass transfer coefficient for this system.varies with F-factor to the 0.72 power while the coefficients for the air-water system are a function of F-factor to the 0.575 power. If the foregoing information is used with Equations (131a) and (156) then the expression for the mass transfer coefficient becomes, kLa = " DL/2 (lPJpL) (157) where i" is a function of weir height and liquid, or possibly more directly a function of slot submergence and JL/PL The variation.of the exponent on the F-factor or gas rate with liquid kinematic viscosity as predicted by Equations (153a) and (155a) cannot be verified by the experimental data. However, it is believed to be significant that the exponent on the F-factor or gas rate was predicted to be in the range of 0.25 to -0.167 and was found to be about zero. The results in Figure 53 indicate that the transition region for the exponent on the kinematic viscosity in Equation (131a) is in the range of 0.05 to 1.0 ft2/hr. It is of interest to see if this range can be predicted by use of the results reported by van Krevelen and Hoftijzer.(88) From the correlation for the characteristics of bubbles

-197in series, the the characteristics become independent of liquid viscosity may be described as follows, Vb db PL 100 (158) I-L g db`nP - 2 (159) Vb2 P By combining these equations with the relationship between flow rate, bubble diameter, and bubble frequency. it is possible to derive an equation which may be used to solve for the kinematic viscosity. 1/2 JPL= 36 (3gQG/1Ns/) (160) where ILJTPL = liquid kinematic viscosity, ft2/hr, At a superficial gas velocity of 6 ft/sec., the value of tI/PL according to Equation (160) is 6.38. Therefore, it appears that the predicted value of IL/PL is too high by a factor of about ten. The most likely reason for this discrepancy is the value of the Reynolds number used in Equation (158). If in place of 100, the value of 1000 for the Reynolds number is used in Equation (158) the predicted value of PJPL then becomes 0.638. The possibility of an error in the bubble frequency being the cause for the discrepancy is unlikely since a factor of ten in the liquid kinematic viscosity would mean a factor of 100 in the frequency because of the square root relationship in Equation (160). It should be pointed out that the foregoing analysis is based on the premise that gas flows through the liquid on the tray in a form approximating a bubble. At low gas rates, i.e., below F-factor equals

-1981.0 this model is probably a very naive representation of a complicated process. The model does serve however to explain on a semi-quantitative basis the -results represented by Equation (131a). At high gas rates or at intermediate gas rates and slot submergence. the gas flows through the liquid. in a nondescript form. In this case, it would be difficult to think of an appropriate shape to use in describing the relationship between the mass transfer coefficient, interfacial area, and rate of surface renewal. In both cases, the times of surface contact are probably randomly distributed, and the model proposed by Danckwerts(22) would most likely describe the mass transfer coefficient better than the model proposed by Higbie31e(42) Similar studies directed toward determining the effects of liq*uid properties on the liquid-phase mass transfer in the two-phase system on a bubble-cap tray are not numerous~ In addition the studies which have been performed have not yielded conclusive.results regarding the effects of liquid properties. The correlation reported by Etherington(27) and represented by Equation (91) includes data from several sources(28O44,64) and is similar to the correlation of the data from the present study.~ However, as mentioned previously, the transition from a small effect of liquid viscosity to a large effect oocurred.at 3 instead of 25 centipoises as found. in the present study. The data used by Etherington covered the range of viscosity from.0.4 to 13 centipoises. Diffusivities were estimated by use of Arnold' s equation but the abnormality factors were unknown. In view of the correlation presented in Figure 51 it! is predicted that the diffusivities for the hydrocarbon.systems used by Etherington were in error considerable~ In addition, Etherington arbitrarily assumed that the

-199liquid diffusivity should be included in the correlation equation to the two-thirds power. Therefore, it is believed the data used by Etherington should be correlated again before a direct comparison is made with the results from the present study. The studies of liquid-phase mass transfer in other types of gas-liquid contacting devices have been more conclusive with respect to the effects of liquid properties than the studies concerned with bubblecap trays. For the case, of wetted-wall columns, Hatta and Katori(41) derived the following equation to describe the mass transfer coefficient, kL= J DL (161) PL BF Z where Z = the length of the wetted surface. r = mean velocity of flow. PLBF Equation (161) has been used to accurately predict experimental data obtained from wetted-wall columns where the time of contact between the gas and liquid was small.(76) Extensive data for packed towers have been obtained by Sherwood and Holloway.(77) Below the gas velocity where loading begins, the H. T. U. s were found to be independent of gas velocity. These data were satisfactorily correlated by use of the following equation, HL= (L)n (L 0.5 162) where a = constant which varied from 80 to 550 for different packings. n = constant which varied between 0.22 and 0.46. L;= liquid rate.

-200It is believed to be significant that in both cases, wetted-wall tower and packed tower, the mass transfer data were correlated by use of liquid diffusivity to the one-half power. In both cases, the time of interfacial contact between the gas and liquid is short and the equations byDanckwerts and Higbie seem to apply. It is logical to expect the same equations to apply for the case of mass transfer on a bubble cap tray. However, it should be pointed out that a recent study by Toor and Marchello(86) indicates that the penetration theory is a limiting case of a more general theory which embraces the penetration and film theories. The study by Toor and Marchello consisted of a mathematical treatment of; the problem of mass transfer to a surface which is periodically renewed. They conclude that at low Schmidt numbers a steady-state gradient is set up very rapidly in any new surface element so that unless the rate of renewal is great enough, steady-state transfer takes place through a film. The rate of mass transfer in this case would be expected to be proportional to the diffusivity. At high Schmidt numbers, Toor and Marchello state that the time necessary to set up the steady gradient is increased and even low rates of surface renewal are sufficient to keep most of the elements from being penetrated. The transfer then follows the penetration theory. For the intermediate cases Toor and Marchello derived equations to indicate the relationship between the rate of mass transfer and the diffusivityo Judging from the results of this study it would not be surprising to find that data from any one type of contacting device might be correlated by diffusivity to the exponents between 0.5 and 1.0 if the physical properties of the system are varied. over a wide range.

CONCLUSIONS Hydraulic Characteristics 1. Liquid density, viscosity, and surface tension were not found to have a significant effect upon the hydraulic characteristics of the tray used in the present study. Data from the literature plus data from the present study were used to determine the effects of the foregoing variables upon froth height, clear liquid height, and gas holdup over the following ranges: viscosity: 0.6 - 97.3 cp density: 0.78 - 2.15 gm/cc surface tension: 21.5 - 70 dynes/cm. Minor variations of the hydraulic characteristics with liquid properties were noted. However, these variations are believed to be within the range of deviations expected with the measurement techniques used in the present and previous studies. 2. The major variables which may affect the hydraulic characteristics of the bubble-cap tray used in the present study are: gas rate, liquid rate, weir height, splash baffle height, and gas density. The hydraulic characteristics at a constant weir height and liquid rate and variable gas rate and gas density may be satisfactorily correlated by use of F-factor, us N. -201

-202Gas-Phase Mass Transfer 1. The coefficients for mass transfer in the gas phase as defined by Equation (23a) were correlated as followse kGV a = 525o.9 0D526 (FFFactor)n kG a = s5230-e9 8) DZ)(123a) PL~-834 (zf zc)0o 28 where n = 0.852 (~L/pL)0~ 238 Phe data from several systems used in the present and previous studies were predicted by use of Equation (123a) with an average absolute percent deviation of 14-59 between the experimental and predicted data. The range of variables represented by the data used in the above correlation are as follows: gas density, PG. 0.0130 - 0.2737 lb/ft3 gas viscosity, G: 0.0351 - o0.0485 lb/ft-hr gas diffusivity, DG~ 0.210- 4.22 ft2/hr liquid density, pL: 0.78 - 2.15 gm/cc liquid viscosity, uL: 0.6 - 24 cp F-factor, Us 0.2 - 1.2 gas holdup,, Zf -"c 2 - 10 inche, liquid rate: 4.6 - 32.0 gpm weir height: 1-1/2 - 3-1/2 inches 2. From the results represented by Equation (123a), the following conclusions are possible: (a) When correcting the mass transfer coefficients for differences in gas properties between two

-203different systems, the use of Schmidt number to the minus one-half power is not justified. (b) The relationship between the mass transfer coefficient and F-factor varies with changes in liquid kinematic viscosity. (c) The mass transfer coefficient decreases as the gas holdup on the tray increase due to increase of weir height at a constant liquid rate or increase of liquid rate at a constant weir height. Liquid& Phase Mass Transfer 1. The complete correlation of the liquid-phase mass transfer coefficients is uncertain because of the questionable diffusivity data available for the carbon dioxide-cyclohexanol system. However, the form of the correlation equation is believed to be as follows: kLa -' DL1/ (IL/PL) (F-Factor)0'58 (131a) where i' and o are functions of I/PL In the range of kinematic viscosity greater than 1.0 ft2/hr, the value of a is minus 0.5. Below the value of 1.0 ft2/hr., a decreases and approaches zero as kinematic viscosity approaches the value of 0.01.: The constant, i'i varies with kinematic viscosity as shown in Figure 53. However, the absolute value of A at any level of kinematic viscosity can not be specified because of the uncertainty in the

-204diffusivity data for the carbon dioxide-cyclohexanol system. The range of variables represented by the data used in the above correlation are as follows: liquid viscosity, PL: 2 - 235 lb/ft-hr liquid density, PL 58 - 62.4 lb/ft3 F-factor, us 6: 0.3 - 2.2 liquid rate 4.6 - 32.0 gpm weir height: 2 - 3-1/2 inches 2. By use of a simple model which assumes that the gas flows through the liquid or froth on the tray in form of bubbles. the correlation between bubble characteristics and liquid properties suggested by Van Krevelen and Hoftijzer(88) and the penetration model proposed by Higbie, the effects of the liquid properties on the mass transfer coefficient were predicted to be in accordance with Equation (131a). This model gives some insight into the effects of liquid properties upon the individual terms, kL and a. The principal conclusion is that the interfacial area and the coefficient are not unrelated terms. In addition, when liquid viscosity is increased, the interfacial area approaches a linear dependence on viscosity but the coefficient approaches a square-root relationship with viscosity so that the comb ined effect is the square root of viscosity. As viscosity is decreased in the range of kinematic viscosity below 1.0 ft2/hr., the interfacial area and consequently kL become independent of viscosity.

Comparison Between Gas- and Liquid-Phase Mass Transfer When the liquid-phase mass transfer coefficient, kLa, was converted to kLa, where the interfacial area is based on the gas holdup per unit volume, the coefficient was found to be practically independent of gas rate. A common factor between the liquid- and gas-phase coefficients is the interfacial area. Therefore, the ratio of the two coefficients may be related to F-factor as follows: kL f( LL, Zf - Zc) (F-Factor)n (163) kL DL PL On the basis of the simple model used to describe kL, it is believed that this coefficient is almost independent of F-factor. It is therefore concluded that the variable coefficient of F-factor which is a function of liquid kinematic viscosity is due to an effect on kG and can be attributed to the variation in circulation or turbulence in the gas phase as liquid. viscosity is varied. Liquid Mixing Experimental point efficiencies were compared with point efficiencies predicted by use of the relationships suggested by Warzel(92), Crozier(21), and Robinson.(72) It is concluded that these relationships may be used to satisfactorily predict point efficiencies when the data are in the range of XEOG below 2.0. In addition, a more critical analysis of the relationships between point and plate efficiency would require data in the range of AEOG above 2.0

REC OMME N DA TIONS Gas-Phase Mass Transfer In the present study, the gas-phase viscosity was not varied over a significant range. Consequently, the effect of this variable on gas-phase mass transfer is unknown. It is recommended that future- studies be performed by-use of systems with widely different gas viscosities. Because of the limited range of gas viscosity which may be covered by gases such as nitrogen, carbon dioxide, helium, etc., the effect of this variable would. have to be studied by -use of distillation systems or vaporization systems where the temperature is varied over a large range, The correlation of the gas-phase mass transfer data as represented by Equation (123a) is not a least-squares fit of the data, This probably accounts for some of the deviation between the predicted and experimental data. It would be of interest to recorrelate the vaporization data by use of a non-linear regression program wherein the correlation equation is the form of Equation (123a), By use of the non-linear regression program confidence limits could be established. for the various coefficients in the correlating equation. In addition, it might be possible to separate the effects of gas holdup and slot submergence on the mass transfer coefficient. The results of the present study indicate that liquid properties significantly affect the gas-phase mass transfer coefficient0 It is believed that the liquid properties should.also affect the energy loss in the gas phase and that there should be some relation between the energy loss and the mass transfer coefficiento It is recommended that in future studies accurate pressure drop measurements across the tray and the -206&

-207bubble caps be made. These data plus an accurate measurement of liquid holdup or clear liquid height could be used to predict the energy loss due to the gas flowing through the froth on the tray. It would then be of interest to compare the rate of mass transfer and the energy loss in the case of the bubble-cap tray with similar data from other contacting devices such as packed towers and wetted-wall towers. Liquid-Phase Mass Transfer The results of the liquid-phase mass transfer study indicates a variable effect of the kinematic viscosity upon the mass transfer coefficient. However, additional investigation of this effect should be performed in the range of 0.1 to 1.0 ft2/hr for the liquid kinematic viscosity. Possibly data in the literature could be used for this investigation. Specifically, Etheringtonns data which covers the range of viscosity between 0.4 to 13 centipoises or 0.022 to 0.72 ft2/hr should be recorrelated using diffusivities determined by use of recent data reported in the literature. In order to completely establish the correlation of the data from the present study, additional data for the diffusivity of carbon.dioxide in cyclohexanol are necessary.

APPENDIX A DESCRIPTION OF APPARATUS The apparatus used in this investigation was designed by personnel of the Chemical and Metallurgical Engineering Department of the University of Michigan. Warzel(92) and Ashby(7) used this equipment during the course of their investigations of plate efficiencieso These investigators have given a detailed description of the equipment. However, the test tray and the flow schemes will be described herein in enough detail to enable another investigator to repeat the same or similar studies. The apparatus was used by Ashby(7) just prior to the present studies and except for a few modifications the equipment is identical to that used by the preceding investigator. The modifications were as follows: 1. Installation of a third blower in parallel with one of the previously installed blowers to increase the range of gas velocity through the test tray. 2. Fabrication and installation of a small heat exchanger between one of the recirculating pumps and the inlet to the test tray to remove heat from the liquid and maintain adiabatic conditions on the tray during the vaporization studies. 3. The rope-type packing glands in both centrifugal pumps (Model 40, Series WS7RD-74 Durcopumps, with 7-1/2 inch open empellers) were replaced by mechanical seals. This was done to eliminate packing gland grease as a -208

I 17,.if — 2 2 1 F F I It~~ Ia -J I I I L I II{ I | -I FRTOP VIEW = i! IDRIll!INE I I N N HOLES ROLL- I + RISERS NTO TRAYS. l, I ~I~i I ll DOWN- I i -I —-- I FRONT VIEW I I I Figure lA. Bubble Cap Plate Layout

'; 2N__'..''Z.:. 9R l s i l -S RS ffi | | -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~................................................ Rs, | l @ | |. l | K | -............................................................................. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:i E E E ~i;;:i~i~2iiii-;s:.- E.:r~l;?i~i~ia:~~ -;-B-.~~~il i...........B,.. -.B.. i -: i. - 9: - ai B:; B ia9:,.,.., 9;, g,,,,;, i,.,................. ii,:-i- 90,9i i l i<, l j 9 i'''B i9 9'"::Biilaa l:X;i, 9:'B;~l - 9 g~~~~~~~~~~~~~~.,... -......... aa.9 a ra B B9R B' BB~~~~~~~~aRBB sB a B a R BBBR,:.a BBBBBa BB R BBB............... - < B B a9 B9 99B Bys; BBB B-B~~~~~~~~~~~~~~~................;^C lggg< aaR gR,.... |.:9. -ii::i-.:j:::::::.,.,:B.-..g.9,g-:; a a: Figure'2a. Removable T~rays,, Showinlg Bubb le - a a Risers for Absorber.

-211possible source of contamination in the liquids being used in the studies. 4. Fabrication and installation of new windows on the front of the column which could be sealed by use of Teflon tape and Epoxy resin, Shell No. 828, materials which were not affected appreciably by the liquids used in the studies. 5. Installation of froth sampling probes above the test tray and liquid sampling devices on the floor of the test tray. Test Tray A general view of the equipment is shown in Figure 2. The test column was designed for five bubble-cap trays but in the present investigation only the first and second trays from the bottom of the column were used. The first tray was used for the test tray, A detailed drawing of this tray is presented in Figure 1A and the detailed dimensions are given in Table VI. The tray contained nine 1-1/2 inch bubble caps arranged on a 2-1/2 inch square pitch. Figure 2A shows two removable trays, one equipped with the caps and the other one with the risers on the removable tray. The outlet weir on the: tray was designed to be adjustable and in the present investigation three different weir heights were used. These were: 1-1/2, 2, and 3-1/2 inch. To eliminate a large hydraulic gradient on the tray an adjustable splash baffle was used. The bottom of the splash baffle in all studies in this investigation was one-half inch above the top of the weir and one-inch in front of the weir.

-212Figure 3A. Position of Probe for Outlet Vapor Sample

-213In Figure 5 the location of the liquid sampling points are shown. Figure 6 shows the liquid sumps below the sampling points. The liquid sumps were made by first cutting the tops from four risers and then welding these into holes in the tray floor and covering the holes with No. 8 mesh wire gauze. In the bottom of the sumps, 1/4inch tubing fittings were welded. Stainless steel tubing (thick walled) was fitted to the 1/4-inch fittings and the tubing was run through compression fittings in the stainless steel panel below the test tray to the outside of the c olumn. By the use of 1/4-inch ells, tees, and valves arrangements were made so that the liquid samples could be taken by use of a hypodermic syringe and needle or by use of a valve. In addition, arrangements were made so that the liquid sampling points on the tray could be used to measure the hydrostatic head of liquid above each point. The tray above the test tray was used to complete the approximation of the boundaries which exist for any tray in a column. It also served the purpose of an entrainment separator and provided a method for obtaining a vapor sample without the entrained droplets. Ashby(7) designed this tray and found that if the vapor sampling device shown in Figure 3A was used, vapor samples devoid of entrainment could be obtained. The caps on this tray were modified by removing the teeth or slots. Ashby's reasoning on this problem was that the entrained droplets from the tray below would impinge on the inside of the caps and flow down the sides. Therefore, if the small sample probe was located in the anmulus between the inside of the cap and the riser, a representative vapor sample could be obtained. It does

-214seem likely that there is a certain probability of a liquid droplet hitting the probe under certain conditions. However, the fact that the data obtained under conditions of high entrainment for certain systems could be correlated with the data for another system (nitrogen-ethylene dibromide) at the same operating conditions but with little or no entrainment would seem to prove this sampling method. This will be discussed further in the Discussion of Results. The liquid sampling probes used to sample the froth are shown in Figure 4 and consisted of a 1/2-inch piece of 3/8-inch stainless steel tubing vielded to tWhv* a 1/8-inch, thin-walled, stainless steel tubing. The end of the 1/8-inch tubing was bent upward so that the piece of 3/8-inch tubing served as a liquid collectorO The 1/8-inch tubing extended through the plate above the test tray where a connection was made to 1/4-inch polyethylene tubing. The polyethylene tubing was taken through compression fittings in the wall of the column to the outside of the column where syringe stoppers were attached to the end of the tubing. Liquid samples were withdrawn through the syringe stoppers by use of a hypodermic syringe and needle. Three sample probes in the froth were used. The horizontal positions of these probes were as follows: one probe was located near the inlet downacomer approximately above the first sample point on the floor of the tray; another probe was positioned above the center row of caps; and the third probe was approximately located above the last sample point on the tray. The vertical positions of these probes were made adjustable by attaching a bracket to the 1/8inch tubing on the second tray and then attaching a single rod to this bracket. The rod was run

-215to the top of the column and through a compression fitting to the outside of the column. By loosening the compression fitting, the rod could be used to adjust the vertical position of the probes on the test tray. Windows The column was originally designed to have glass panels on the front side of each of the five trays. Thus the hydraulic action of the- column could be observed when the column was in operation. Ashby7 used Tygon tubing and partially-polymerized styrene to make a seal between the glass panels and the face of the column. Since it was planned to use ethylene dibromide, a very good solvent for many materials, in the present study, it was realized that a different type of material would have to be used to form a seal. Teflon was found to be resistant to the solvent power of ethylene dibromide. But the Teflon was not resilient enough to form a seal when used with the glass panels as originally designed. Therefore, a new design of the windows was made. In this design, a 3/8-inch thick, stainless steel, plate was used to make a frame which would provide a flat surface to make a seal at the face of the column. One of the carbon steel window frames used by Ashby(7) and Warzel(92) was welded to one face of the stainless steel frame. The open area in the stainless steel frame was smaller than that in the former frame. Thus when the carbon steel frame was welded to the stainless steel frame, the difference in the areas of the two openings was represented by a small rectangular anrnulus around the inside of the carbon steel frame. Teflon tape and Epoxy resin (Shell No. 828) was used to form a seal between

-216a piece of 2-ply safety glass and the surface of the small rectangular annulus. The whole assembly was bolted to the face of the column. The seal between the stainless steel frame and the finished (ground) surface of the column face was also made by use of Teflon tape and Epoxy resin (Shell No. 828). -A good seal was obtained in most cases by first painting the Teflon tape and the areas of contact on the column and the frame with Epoxy resin. The window was then put in place on the column and bolt pressure was used to distribute the resin smoothly After the resin hardened, the tray was ready for operation. Windows of the type described above were used on the test tray and the first tray above the test tray. Stainless steel panels, 1/8einch thick, were used on the other trays of the column. The panels were of the same over-all dimensions as the carbon steel frames used by Ashby and Warzel. In fact, these frames were used to back-up the thin stainless-steel panel and the two pieces were bolted to the column. The seal between the stainless steel panel and the face of the columna was also made by use of Teflon tape and Epoxy resin. Arrangemests for Liquid Sampling, Hydrostatic Head and Pressure Drop Measurement The hydrostatic head was measured at four points on the tray floor by use of the liquid sampling points shonm in Figure 4. In Figure 4A a schematic diagram is presented for the arrangement of the valvies and tees on the outside of the column which made it possible to use these points for obtaining liquid samples and for measuring the hydrostatic head. The liquid line from each of the four points on the tray were;connected to a tee fittang (No. 1) ona the outside of the

To Pressure Probe Above Test Tray Hydrostatic Head I Y4-Inch Tees (No. 1) Connected to Lines From / Floor IRvel oIf Liquid Sampling Pts. e Ta 9~~~~ Syringe Stoppers' iTubin Bokeanei Hk Valve s ~ ~ ~ II'n -n ii IValve Fiue4.Apaau sd omanLqudSmls n esr Hdottc haaFo Potonrylo Bott le~~~~~~~~~~~~~~~~~ O Polyethylene O~~~~~~~~ Connected to Lines From Floor level -ofor

-218column, One side of each tee fittin.g (Noo 1) was used to connect a line to tee fitting, No9 2, and other side of Noo I was used to connect a line to one side of a Hoke, quick-opening valve. The other side of the Hoke valve was connected to glass tubing mounted on a backboard9 The glass tubing was mounted in a vertical position and the top of each tube was connected to a pressure probe above the test tray~ A scale in 0.2-inch graduations on the backboard was used to measure the hydrostatic head above each point on the tray. The zero;point of the scale on the backboard was carefully adjusted to the tray floor leveL Therefore, the liquid level in each glass tube as indicated by the stale was equal to the hydrostatic head at the- particular point on the test ztray. The No. 2 tee fittings were used to obtain liquid samples by use of a hypodermic syringe or a small Hoke valve. A syringe stopper over -one side of the fitting was used to obtain the samples by use of a hypodermic syringe. The other side of the tee fitting was connected to a Hoke needle valve~ A 1/4-inch line from the needle valve was fitted to a piece of 1/4-in.ch, staimless-steel, tubing9 This piece of tubing was ground on one end to form a sharp edge which could be used to puncture the syringe stopper in the sample bottles. Therefore, the tubing could be used to obtain liquid samples by first inserting it into the sample bottle and then closing the quickopening vaLve in the line to the glass tubing and opening the small needle valve. The pressure above the test tray was determined by use of a 1/4-inch probe connected to a 304irch mercury-filled manometer. The pressue drop across the test plate was measured by a water-filled mnop meter.

-219Gas Recirculation - Vaporization Studies The flow diagrams for the vaporization and absorption studies are shown schematically in Figures 5A and 6A, respectively. In the vaporization studies, the gas plus the cyclohexanol or ethylene dibromide vapors and entrained droplets passed to an entrainment separator where part of the entrained droplets were removed. From the entrainment separator the gas and vapors plus any entrainment remaining in the gas flowed to blower #1. On the suction side of the blower nitrogen was added to make up for leakage and maintain the pressure in the system. Also an orifice plate with one, two, three, or five, 5/8-inch holes was used in a pipe flange on, the suction side of the blower to maintain a positive pressure in the test column, At low flow rates the orifice plate with oe hole was used but as the flow rate was increased the nixnber of holes ia the orifice plate was increased so that the capacity of the blower was not decreased and so that the pressure in the test column did not become greater than that considered safe for the glass windows. After leaving blower #1 the gas entered the bottom of a 12-inch diameter, 4-plate, sieve tray column where it contacted either cyclohexanol or ethylene dibromide, the liquid being used at the time. The liquid in the sieve tray column was maintained at a temperature approaching that of the tap water in the Chemical and Metallurgical Engineering Departmental Laboratories. Thus part of the vapors in the gas was cdensed and the gas returned to the test tray with the vapor content appreciably reduced. Before returning to the test tray, the gas was metered by use of a Fischer and Porter Flowrator (Size 12, Tube No. 12-LL25, Serial No. D8-1609, Figure No. 26PE, Chemical and

GAS ROTAMETERS GR F P, SERIAL NO. 8- 1609. GR SCHUTTE — KOERTING, SERIAL NO.44176. LQUID ROTAMETERS LR, F&P, SERIAL NO. W70- 4024.1 LR2, F&P, SEWIAL NO 5601 01038 1. HEAT-EXCHANGERS. SSr W CONTROL VALVES. ATOR DEHUIDIFER BL OWER 1) 4 SIEVE PLATES 3RD FLOOR OF LABORATORY._ LA 2_ > STEAY L TEST COLUMN STEAM LR LR2 GR3 TO TlEST TRAY "PANCAKE* PRESSUR REGULATOR N2-SUPPLY WACER 2ND FLOOR OF LABORTORY,,!j DURCO PWUP DURCO PUMP Figure 5-A. Flov Diagram for Vaporization Studies

GASLRWAMEIE GRI FS P, SERIAL NO. 06- 2445 OR DO0-1609. TO GR2 FOP, SERIAL NO. DO- 1611. VENT. GR3 SCHUTTE- KOJERTING, SERIAL NO. 44176. -HEAT AIR~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~TR SLOWER S~~~~~~~~~~LOWER 2BLOWERI 12-NHDA 3RD FLOOR OFLABORATORY 4__SIEVE_____ -J.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~\ TO ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~O DRAIN~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~T TEST COLUMN~DRC PM ORC PM Fiue6A.FR iaga or2 AbsrptonStuiR

-222Metallurgical Engineering Department No o C17Dm200). Scale graduations were in increments of 2 cubic feet per minute and the tube was calibrated by the manufacturer from 0 to 200 cubic feet per minute for a gas of 0~877 gravity at 147 psia and 60~F. Warzel(92) checked the calibration of this flowrator and found the scale calibration to be sufficiently accurate for this type of study. In order to maintain constant conditions at the test tray, an 8-inch diameter Ross heat exchanger (Type "SSCF", No. 804) was used to heat the dehumidified gas after being meteredo The heat in the exchanger was supplied by condensing steam at atmospheric pressure~ Liquid Recirculation Vaporizat ion Studies Two separate liquid recirculating streams were used in the vaporization studies6 These were: (1) The liquid recirculated to the test tray and (2) the liquid recirculated to the dehumidifiero The liquid left the test column at the bottom and was recirculated by a centrifugal pump (Model. 40, Series WS7RD-84 Durcopump, with 7-T/2 inch open impeller and mechanical seal, Serial No. DU 315l/52). This mechanical seal is manufactured by the Durametallic Corporation, Kalamazoo, Michigano Between the pump and the test trayr, the liquid was metered by use of a Fischer and Porter Flowrator (Size 8, Series 700, Serial Noo W704024/l., Tube No. B9%27l0/70-G and Float No. BSVT93)- A portion of the liquid from the pump was by-passed to a small heat exchanger where heat was removed by cooling water. Considerable heat was gained by the liquid in the pump. This was especially true for the cyclohe-xa ol at a high viscosity, ioe leow temperature Th.us,

-223the small heat exchanger was used to aid in the maintenance of adiabatic conditions on the test tray. The liquid to the dehumidifier was also circulated by a DTurco centrifugal pump equipped with a mechanical seal A Ross heat exchanger similar to that used in the gas recirculation system was used to remove heat from the liquid and maintain the lower temperature in the dehumidifier. Gas Recirculation - Absorption Studies The gas recirculation system in the absorption studies was somewhat simpler than in the- case of the vaporization studies. The mixture of carbon dioxide, air, cyclohexarol vapors and cyclohexanol entrained droplets flowed from the test colwm.to the-.entrainment separator and then to either one:or two blowers. When two blowers were used in parallel to increase the range of gas velocity at the test tray, part of the gas from the entrainment separator flowed to Blower #1 and the remainder to Blower #2. Orifice plates in a pipe flange on the suction.side of each blower were used to regulate the level.of pressure in the test column Carbon dioxide gas was added to the system on the suction side of Blower #1. The- rate of carbon dioxide fed to the system was dictated by the absorption rate on the test tray and by the leakage rate at the pressure in the system. The addition of carbon dioxide was controlled by a "pancake" pressure regulator which is distributed by the Mathieson Company, Joliet, Illinois. After being compressed at the blowers, the gas was metered. One of two rotameters was used to meter the gas: (1) When the absorption studies were being conducted in the 1w range of velocity a Fischer

-224and Porter Flowrator, Size 12, Tube No. 12-LL25, Serial No. D8-1609, Figure No. 26P-E, Chemical and Metallurgical Engineering Department No. C17-200, was used and (2) For the high range of velocity, a Fischer and Porter Flawrator, Size 12, Tube No. 12LL-25, Serial No. D6-2445, Figure No. 26'P-E, Chemical and Metallurgical Engineering Department No. C17-235, with a type 347 stainless steel float, No. D8-1617, was used. The latter rotameter was calibrated by use of a Roots-connersville rotary meter using a natural gas at the "J" Station of the Michigan Consolidated Gas Company, Detroit, Michigan. The calibration data are presented in Table IID in Appendix D. The gas mixture passing through the flowrator was in most cases at a higher temperature than the temperature desired at the test tray. This is because of the heat of compression and the heat gained from the blowers which usually were hot to the "touch" when running. Consequently, the gas stream had to be -cooled in the Ross heat exchanger used in the vaporization studies. The cooling water in this case was tap water which had been heated by steam in an auxiliary heat exchanger. Water from the same source was used to maintain the temperature of the- liquid to the test tray. Liquid Recirculation - Absorption Studies In the absorption studies, one of the centrifugal pumps was used to feed eyclohexanol to the test column and the other one was used to pump the liquid from the bottom of the test column to the top of the desorber (the dehumidifier in the vaporization studies). In the desorber the liquid was contacted with air from the laboratory and the

-225carbon dioxide was desorbed in order to maintain a low concentration in the feed to the test tray. Before rereturning to the test trays the liquid was metered in a Fischer and Porter Flowrator (Size:8, Series No. 700,, SerSial No. 5601D1038B1, Tube No. B9-27-10/70-G and Float No. BSVT-93). The calibration data for this meter are given in Table I.IID,. Appendix D. After being metered, the liquid flowed tchrough a Ross heat exchanger where the liquid was either heated or coled befoe returning to the test tray.

APPENDIX B LABORATORY TECHNIQUES Vaporization Studies The two factors which were important in obtaining reproducible and accurate data in the vaporization studies were as follows: (1) Adjusting the temperature of the gas and liquid at the test tray to adiabatic conditions and (2) obtaining representative samples of the gas entering and leaving the tray. Adiabatic Conditions Theoretically, adiabatic conditions are obtained by setting the inlet gas temperature at the value desired and then controlling the heat added to, or removed from, the liquid so that the temperature of liquid entering the tray is equal to the temperature of the liquid leaving the tray. When the temperature of the liquid ceases to change as it flows across the tray, adiabatic conditions have been established This is essentially the procedure used in the present studies. However, in the present case the liquid temperature, ide., the adiabatic saturation temperature of the gas, was.the known factor and the gas temperature was adjusted to the point where the liquid temperature did not change significantly as it flowed across the tray. In order to eliminate the time required to search for the adiabatic conditions by trial and error adjustment of the temperatures at the test tray, humidity charts for cyclohexanol and ethylene dibromide in nitrogen were prepared. These charts are presented in Figures 1lB and 2B. The following equations were used in the preparation of these charts, -226

-227T = (as X)XTas Ta s (1B) hG (T - k (2B) hG _C(IIGPGDG 2/3 (3B)2 kG MG P P p/k (3B) where 3-= humidity of the gas, lb vapor/lb dry nitrogen 3+as = humidity of saturated gas at Tas, lb vapor/lb dry nitrogen Xjw = saturated humidity at the wet bulb temperature Tw. Tas = adiabatic saturation temperature, OF Tw = wet bulb temperature, OF T = gas temperature, QF XTas = latent heat of vaporization at the temperature Tas, BTU/lb Tw = latent heat of vaporization at the temperature Tw, BTU/lb hG = heat transfer coefficient, BTU/hr/OF/ft2 k = mass transfer coefficient, lb moles/hr/ft2/atm. G P = total pressure, atm.. MG = molecular weight of gas, lb/lb mole Cp = specific heat of gas, BTU/lb/mole/OF S = specific humid heat capacity of the gas, BTU/lb dry gas/ F.'G/PGDG = Schmidt Number Cpl~kq = Prandtl Number

0.06'4/ ~A z J w 0.05 0 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~.~~~~ 0. 0 6, 2~~~~~~~~~~~~~~~~~~~ A.0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0 0~~~~~ Op 2 0.03 0 U' 70so9 10 0 10 3r10o5 oHi a t t y rS 010.b L.0O 0.0 0o 70 80 90100 110 20 130 10 150 16 O ~ ~ ~ ~ ~ ~ ~ ~ ~ EPRTR 0F, Figure -B HuiiyCatfor h irgnCcoeao ytmPeae by C..W

us ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 0.36_ 324 0 U~~~~~~~~~~~~~~~~~~~~~ cd Acb r A 0 016.a,, ODS ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~~~~~P 0.12~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 0~ -04-:es, "~~~~ Woo 30 40 50 60 70 80 90 I00 110 120 130 140 150 160 r7o 180 TEMPERATURE OF Figure 2-B. Humidity Chart for the Nitrogen - Ethylene Dibromide System Prepared by C. -H. Wu

-230The vapor pressure data for cyclohexanol and ethylene dibromide which were used in the preparation of these charts are presented in Tables IVE and VE. The term, hG/kGMGP, was estimated by Equation (3B) wherein the physical properties, k4, Cp, Ft, and X were taken from the literature and DG was calculated using the equation developed by Wilke and Lee (97) For cyclohexanol in nitrogen, the term, hG/kGMGP, was estimated to be 0.540, and for ethylene dibromide in nitrogen, hG/kGMGP, was found to be 0.505. By knowing the liquid temperature desired and then estimating the inlet gas humidity, the inlet gas temperature was determined by use of the adiabatic cooling curves in Figure 1B or Figure 2B. Usually the inlet humidity of the gas was estimated from a preceding run. The adiabatic conditions estimated by use of the humidity charts were in most cases very good estimates judging from the difference between the inlet and outlet liquid temperatures. Occasionally these temperatures differed by as much as O.30C but this difference could be decreased by changing the inlet gas temperature. In the majority of the runs made, the difference between the inlet and outlet liquid temperatures did not differ by more than 0,10C during sampling. There were exceptions however where this difference was as high as 0.2~Co Some difficulty was experienced in obtaining adiabatic conditions at high gas rates but this was thought to be due to inadequate controls on the equipment. The liquid temperatures were measured by calibrated mercury-in-glass thermometers with O.1~C scale graduations from -1 to 51~C. The gas temperature belw the tray was measured by a calibrated mercuryoin-glass thzermometer with O.10C scale graduations from -1 to 1010C.

-231Sampling The amount of organic vapor in the inlet and outlet gas was determined by withdrawing the samples from the column through heated sample lines into UTtubes filled with glass beads and immersed in ice baths. Three U-tubes in series were used for each sample. Because both cyclohexanol and ethylene dibromide have freezing points above O0C, the first U-tube in the series of three was immersed in a chilled water bath. If too much ice was added to this bath a large fraction of the total vapor in the gas sample was condensed in this U-tube and the liquid (cyclohexanol or ethylene dibromide) would freeze and block thE flow of gas from the column. The second and third U-tubes in the series were immersed in "dry" ice-acetone baths. In these tubes the vapors were condensed and frozen. Thus the vapor content of the gas leaving the last tube approached the concentration corresponding to saturation at the vapor pressure of cyclohexanol or ethylene dibromide crystals. In all runs, less than 0.5 percent of the total vapor in the samples was collected in the third U-tube and in sme cases, particularly for the outlet gas samples,, this was as low as 0.l percent. These results were thought to be sufficient evidence that more than three U-tubes in series were not necessary for sufficient accuracy in the determination of the vapor in the gas samples. The gas after leaving the U-tubes was saturated with water vapor in a bubbler and then metered by use of wet —test meterse These meters were calibrated in the CM-16 Laboratory in the Chemical and Metallurgical Engineering Department~ The calibration data are given in Table IV-D.

-232The weight of vapor condensed in the U-tubes was determined by weighing the U-tubes before and after sampling. A Christian-Becker projectomatic analytical balance and class S, stainless steel, weights were used to weigh the tubes before and after sampling. Ashby checked the weights against weights which had been calibrated by the Bureau of Standards and found that they were accurate within + 0.2mgo In preparing the U-tubes for the samples, air was passed through the tubes until the cyclohexanol or ethylene dfibromide from the preceding samples was completely evaporated. The glassground, tapered, stopcocks were then cleaned and a lowtemperatlure stopcock grease was used to seal the tubes. The external surface of each tube was then cleaned by wiping with clean cloths. The tubes were ready for samples after weighing Absorption Studies In the vaporization studies, the mass transfer or plate efficiency was deternined by analyzing the composition of the gas entering and leaving the test tray. There are two principle reasons why this experimental technique could not be used for the study of carbon-dioxide absorption by cyclohexanol. These are: (1) The change in the gas composition across the test tray was not great enough to be measured accurately by available methods, and (2) The ccncentration of carbon dioxide in the liquid varied along the length of the tray. Therefore, liquid samples on the test tray were required to determine an average comcentration to be used in the mathematiceal relationship between overall plate efficiency and point efficiency. Even if it had been possible

-233to'measure the change in gas composition across the test tray, the liquid samples on the tray would have been required in addition to calculate the point efficiency using the measured over-all plate efficiency. Therefore, for each run the following samples were obtained: (1) A sample of the gas above the tray which could be used to determine the equilibrium concentration at the temperature and pressure on the test tray, (2) Samples of the liquid in the entrance and exit downcomers, and (3) Samples of the liquid at four different points on the floor of the te-st tray. Test Conditions In order to eliminate the possibility of cooling or heating the gas-liquid interface, the temperature of the gas entering the tray was adjusted so that adiabatic conditions existed. Thus the equilibrium liquid concentration was determined by the bulk liquid temperature. The gas was recirculated to the test tray after being compressed in the blower and heated or cooled in an exchanger. Since the gas was not dehumidified except for some candensation in the circuit, its cyclohexanol vapor content was approximately equal to the saturation composition corresponding to the liquid temperature on the test tray. Therefore, the gas and liquid temperatures at the test tray were approximately equal. The difference in liquid temperature at the entrance and exit of the tray were never greater than 0.2~C when samples were being obtained.

-234The liquid samples taken during the studies at 3-1/2 inch weir height were obtained by use of a calibrated hypodermic syringe. The lines from each sample point an the tray was fitted with a syringe stopper. A liquid sample was obtained as follows: The hypodermic needle was pushed thxough the syringe stopper and the plunger of the syringe was slowly withdrawn to fill the syringe. If the vacuum created by the withdrawal of the plunger became too great the liquid began to degas or air leaked into the line through the hole in the stopper. Therefore, the rate at which the sample could be obtained, was necessarily slow in order to avoid loss of carbon dioxide in the sample or contamination by air leakage. The sample line and tfhe syringe were flushed by liquid from each point by withdrawing 50 to 100 cc of liquid and rejecting it. A small volume of air from the needle was usually trapped in the syringe as sampling began. This was rejected,as the liquid volume in the syringe was being adjusted to the desired volume. Samples taken from the tray floor by use of the syringe were completely devoid of entrained gas bubbles. A few bubbles were obtained in the samples taken in the inlet and outlet diwncomers and in the saples from the froth. These were rejected as the liquid volume in the syringe was adjusted. The use of hypodermic syringe to obtain liquid samples proved too time consuming. Also, a health hazard was created by skin contact with the cyclohexanol and by breathing the cyclohexanol vapors. Therefore, arrangements were made so that samples could be taken through

-235needle valves directly into the sample bottles, This arrangement is shown in Figure 4A. The volume of the liquid sample was determined by weighing the sample bottle on a 3000 gram precision balance before and after sampling, A six-inch piece of 1/4-inch, stainless steel, tubing attached to the polyethylene line from each needle valve was pushed through the syringe stopper in the top of the sample bottle. The 1/4inch tubing was long enough to reach the bottom of the sample bottle so that the tip could be immersed in the barium hydroxide solution. The sample bottle ccmtained a magnetic stirring bar so that a magnetic stirrer could be used to mix the cyclohexanol with the barium hydroxide solution as the sample was being withdrawn. A blank sample was used to indicate the approximate sample volume desired. In order to compare this method of sampling with the method using the hypodermic syringe, each method was used to take cosecutive samples from the same sample point during several runs. The cocentration of carbon dioxide in these samples are compared in Table I-B. With exception of the samples at a concentration of about 0.5 x 10'5 gm-mol/cc, the concentrations in the samples taken by the two methods agreed within -5 percent. The gas sample from above the test tray was obtained as follows: A 250-millimeter gas sample bulb with stopcocks on both ends was first evacuated to less than 0.2 inches of mercury. This bulb was then attached to the sample line from the test tray. An ice trap in the line was used to remove cyclohexanol vapors. Before the liquid sampling was started, flushing of the gas sample bulb was started by first opening the stopcock Con the end of the bulb nearest to the connection to the sample lime and thren opening;the stopcock on the other end.

-236The bulb was flushed during the time that liquid samples were being taken. This was usually 15 to 20 minutes.,TABIE I-B COMPARISON OF SAMPLING METHODS Sample Taken by Syringe Sample Taken Through Valve Carbon Dioxide Cone., gmnmol/cc x 105 0.530 0 482 0.982 1.030 1.475 1.425 1, 560 1.552 1.650 1.700 2.105 2.080 2.130 2.125 2. 245 2.250 Preparatin of Sample Bottles The sample bottles used to receive the liquid samples from the test tray were 8-ounce Boston round bottles made by the LibbyOwens Glass Company. Syringe stoppers were used to seal the top of each bottle. In preparing the bottles for samples, the procedure was as follows: (1) The bottles were flushed with nitrogen and then sealed with syringe stoppers. (2) 50 ml. of 0.1N barium hydroxide was added to each

-237bottle directly from a 50 ml. calibrated, Kimble EXAX,, buretteo (3) When the samples were taken by use of needle valves in the sample lines, a stirring bar was added to each bottle. (4) The bottles were finally weighed on a 3000Qgram analytical balance. Analyt ial Te hnique A 0,1N barium hydroxide solution was used to precipitate the carbon dioxide in barium carbonate when the liquid samples were added to the sample bottles. An excess of barium hydroxide was contained in each bottle. The amount of carbon dioxide in the liquid samples was determined by titratio of the excess barium hydroxide with 0.1N hydro:chloric acid.: The volume of O.lN barium hydroxide added to each sample bottle was 50 ml. in all cases. In order to minimize the errors which were inherent in the sampling methods and in the titration, the size of the liquid sample was as large as could be conveniently handled in the 8-ounce bottles. Usually the sample size was near 100 cc. Precautions were taken to avoid contact of the samples with air. This was accomplished by titrating the excess barium hydroxide directly in the sample bottles. The tip of the acid burette (Kimble NORMAX No. 8275 or Kimble EMAX) was pushed through the syringe stopper and the acid was added as the contents of the bottle were agitated by use of a magnetic stirring bar and a magnetic stirrer. A 10 mi. sample of saturated barium chloride was added to each sample to suppress the solubility of th- barim carbnate whern the samples were titrated.

-238Cyclohexanol is not miscible with water. Therefore the sample bottles contained two phases. -Ir tMese two nhasFps -rrp mTnr A by the magnetic stirrer, a pink opulescent mixture was formed when the indicator, phenolphthalein, was added. As the O.1N hydrochloric acid was added, the mlxture became white. When the stirrer was stopped, the two phases separated but the aqueos phase was still pink in cases where the end point had not been reached or passedo The titration was continued from this point by adding tw o t three drops of acid solution, stirring the contents for approximately thirty seconds, and then allowing the two phases to separateo This procedure was continued until the pink color of the aqueous phase disappeared. In order to check the accuracy of the titrometric method, synthetic samples were prepared by accurately weighing samples of sodium carbonate which were then added to a sample bottle containing 50 ml. of 01lN barium hydroxide and 100 ml. of cyclohexanol which had been degasseda by refluxing'under a vacuum, The sodium carbonate reacted with the barium hydroxide according to the following equationNa2C03 + Ba(OH)2 BaC3 + 2aOH Thus the contents of the bottle were-identical to that of the samples fromthe test tray except for the presence of the sodium -iOns/.~ The synhetaic samples -were titrated with O.1 H Cl, anld the milliequivalents of acid were compared with the'milliequivale nts of barium hydroxide added initially. TIhe two values should have bee equal if the ti. tration method was perfect. The average of.the diffe-rences between

_239the milliequivalents of barium hydroxide and acid for several samples was about 0.084 milliequivalents. This value represents about 1.8 percent of the milliequivalents of barium hydroxide added to the sample bottles originally. In an attempt to estimate the error in the determination of the milliequivalents of carbon dioxide in the samples from the test tray, it was assumed that the same percentage of error, 1.8, could be applied to the barium hydroxide in the samples before titration. The amount of barium hydroxide remaining varied depending on whether the sample was taken near the inlet or the outlet of the tray. Normally this variation was between 1.5 and 3.5 milliequivalents. Since the milliequivalents of barium hydroxide used in each sample bottle was about 5.0, the error in the milliequivalents of carbon dioxide varied between 0.027 and 0.063 or 0*77 and 4.2 percent, respectively. This error is significantly reduced when it is considered that the liquid sample size was 100 cc. In concentration units of lb-mol/ft3, an error of 0.063 milliequivalents in the carbon dioxide determination represents an absolute error of 1.96 x 10-5 lb mol/ft3 for a 100 cc liquid sample. This would be the absolute error for the sample at the tray inlet where the concentration of carbon dioxide was 30 to 40 x 10-5 lb-mol/ ft3. In other words, the error in the determination of the inlet sample concentration was of the order of 5-6 percent. For the outlet samples, the absolute error was of the order of 0.85 x 10O5 lb-mol/ft3 or approximately 0.5 to 1.0 percent on the average. A more detailed evaluatijo of the experimental error for the absorption studies is presented in Appendix C.

APPENDIX C ESTIMATE OF EXPERIMENTAL ERROR Vaporization Studies An accurate estimate of the absolute error in experimental data of the type obtained in these studies is almost impossible,. However, it is important in the interpretation of the data to estimate the experimental error as well as possible.. This is particularly important in experimental studies where the error may vary with the conditions or the materials used in the studies. Such is the case in the present investigation, i.e., the percent of error in the value of EG or NG may vary depending on the liquid and conditions used in the vaporization studies. Ashby(7), who studied the vaporization of several liquids, has presented a summary of the types of errors to be expected and has estimated the percentage of error in NG for each of these inac= curacies. Ashby concludes that the error in NG can be considerable (20-30%) if the estimated errors for the individual sources of errors existed and were additives. The probability that the errors are additive and not partially compensating, is probably small. Ashby divided the errors into the following three categories: (1) errors in the determination of the correct equilibrium composition for the gas on the test tray, (2) errors in the determination of the correct inlet and outlet compositions of the gas, and (3) miscellaneous errors. Since approximately the same tec _hique was used in the present investigation, the foregoing categories are applicable. However, ~it should be realized that each of these categories includes several -240

-241types of error. These are the errors due to inaccuracies in calibration data, the errors in funadamental data such as physical properties, etc., and random errors which depend on the operator or observer as well as the technique used in taking the data and operating the equipment. The sources of possible:errors in each of the factors, y, yln and y*, are listed in the following tables, Possible Errors in the Gas Equilibrium Composition, y* 1. Measurement of the average bulk liquid temperature on the test tray. 2. Assumption that the surface temperature of the liquid was equal to its bulk temperature~ 3. Vapor pressure data. 4. Influence of impurities on the vapor pressure of the liquid. 5. Measurement of the pressure above the test tray. Possible Errors in the Inlet and Outlet Gas Compositions 1. Weighing of the U-tubes or drying tubes. 2. Correction for change in weight due to atmospheric changes in the balance room. 3, Failure to remove all vapor from the gas samples. 4. Inaccuracy of the wet test meters. In order to show how the percentage of error in EG and NG may vary depending on the system being studied, Table IC has been prepared wherein one percent of error in y*u, y0, and y,. was assumed. The following equations were used in the calculations of percentage

-242errors in EG and NG 6EG ~EG (EG dE - dy +'yl + dy* (1C) 6y — yl 6Y where EG = Y* - Yl 1 (dy - dyl + EG dyl EG dy*) (2) EY* - Yl For the maximum error, dEG 1 (yE + Yel + EGyll1 + EGYl + E*) (3C) (YE Y)C) where, = fractional error dNG - E (4c) dNG where NG = - In (l'- EG) (5C)'dNG =EG (6c) The results in Table IC indicate the variation Of experimental error with the srystem studied and explains why the scatteor oreproducibility of the data varied where the same technique Vwas used for all systems. kctually to obtain the same repr6dducibility in the data for the several different systems, the technique should have been varied somewhat. It was necessary to do this in the case Of cyclohexanol. n order to approach the reproducibility in EG wbich Ashby had obtained in his studies, the volume of gas samples was increased to 2 cubic feet. Ashby took gas samples of 0.5 cubic' feet or less. In the ethylene dibromide vaporization studies, good reproducibility of the data was obtained in most cases by takinfg gas samples of 1 cubic feet. On the basis of the results in Table IC the increase in the reproducibility

-24L3_ TABLE I-C PERCENTAGE ERROR IN EG AND NG RESULTING FROM l% ERRORS IN y*, y1, and y System % Error in EG % Error in NG He-Water*' 4.1 13.0 Air-Water* 3, 7 17.5 Freon 12-Water* 2.8 15.2 He-i C4H90OH* 3 7 9 5 N2-i C4H9OH* 3.52 9.8 He-MIBK* 3.8 10.4 N2-Cyclohexanol (EG = 0.55) 6.6 10.2 N2-C2H4Br2(EG = 0. 5) 3.6 5.2 FSystem studied by Ashby of the efficiency data with increase in sample volume can be explaineda by the increase: in the reproducibility of determinaing y and Y1l But why the reproducibility in y and yl increased as the sample volume increased is not known. It would seem that somewhere in the technique of taking the samples or weighing the samples a random error in the amount of cond~ensate was introduced but was minimized by taking a larger sample volume. However, this would mean that the absolute value of the error did not increase with the volume of ga.s sample. This type of error could be introduced by improper handling or treatment of the condensate sample tubes. Another possibility is that there could have been fluctuations in the composition of the gas to or from the test tray. However, since the minimum sampling

-244time was 15 minutes, the frequency of the fluctuations would have had to be greater than 1 cycle per 15 minutes in order to show up in the analysis of the gas entering and leaving the test tray. TABLE II-C RELATIVE MAGNITUDE OF TEE TERMS IN EQUATION 3C Gas Sample Run No. Volumes, Ft3 y y y* y*-y EGO% H-14-A 1.0 0.002291 0.003115 000413 0.001839 44.93 H-14-B 1.0 0.002197 0.003153 0.00412 0.001923 49-71 H-12-A 1.0 0.00234 0. 00282 0. 00417 0.00183 26.20 H-12-B 1.0 0. 002008 0. 002813 0.00413 0.002122 37.90 H-13 -A 1.0 0.002227. 00309 0 o 004066 O 0 001839 46 90 H-13 -B 1 0 0.002260 0.00280 0.00408 0.00182 29.60 H-15 -A 1.0 0,.002184 0o 003440 0.00416 0-001976 63o60 H-17-A 1.0 0.001920 0.-002756 0o.00351 0.001590 52.60 H1-17-B 1.0 0. 00191 0.00293 0.003526 o0o001616 63.10 H-21-A 1.0 0.001891.0 0314 o.003990 0.002099 59.50 H-21-B 1.0 0. 001828 0. 00319 0.00401 0.002182 62..50 H -11-A 1 0 04 002436 0.00315 0.00412 0.001684 42.30 H-11-B 1.0 0.002428 0.00300 0.00408, 0o001652 34.50 H-10-A 1.0 0.002420- 0.00312 0. 00413 0O 00171 41.00 H-10.-B 1.0 0. 00258 0.00334 0 00429 0. 00171 44. 40 H-37-A 2.0 0. 001526 0.003095 0. 004081 0.002555 61.41 H-37-B 2.0 0.001505 0..003057 0.004054 0.002549 60.89 H-38-A 2.0 0.001371 0.002960 0.004151 0.00278 57.16 H-38-B 2.0 0. 001348 0o002931 0.004151 0.002803 56.47 1-39-A 2.0 0.001498 0.003207 0,004162 0OO2664 64.15 H-39-B 2.0 0.001509 0,003197 0.004184 0002675 63.10 *Runs A and B are duplicates.

-245Table II-C has been prepared to show the relative magnitude of the terms in Equation 3C and to show the improvement in reproducibility by increase of gas sample size. Absorption Studies The experimental errors in the efficiency and mass transfer data for the absorption studies are more uncertain than for similar data in the vaporization studies. This is primarily due to the uncertainties in method of obtaining liquid samples and the method of analyzing for the carbon dioxide in the samples. The possible sources of error in the efficiency are those associated with the concentrations, xo, xi and x* and are tabulated below. Possible Sources of Error in the Liquid Concentrations, xo and xi 1. Obtaining samples of the liquid from the two-phase system. The factors involved here are: (a) Withdrawing liquid samples devoid of gas bubbles from the system and (b) Transferring the samples to the sample bottles without loss of the carbon dioxide in the sampleso 2. Determination of the volume of liquid sample withdrawn from the tray. 3. Analysis of the amount of carbon: dioxide in the liquid samples.

-246Possible Sources of Error in the Equilibrium Liquid Concentration, x * 1. Analysis of the amount of carbon dioxide in the gas sample from above the test tray. 2. Measurement of the pressure above the tray. 3. Measurement of the bulk liquid temperature. 4. The possibility that the gas-liquid interface can be cooled or heated and not necessarily equal to the bulk liquid temperature. 5. Solubility data for carbon dioxide in cyclohexanol. In the experimental procedure, every effort was made to minimize each of the errors in the foregoing tabulation. In order to obtain an accurate estimate of the maximum possible error, a detailed study of each of the variables would be required.- Therefore, for the present evaluation only the effect of the error in the analyses of carbon dioxide will be considered. The percentage error in the plate efficiency, EML, may be determined by use of Equation (12C) which was derived as follovs: adE,=; dxo + aL x (70) ~~~~~~~~aEML 1= ~~(8C) 0 o 1 1~~~~~~~~~~~(c

-247aE,- ( x.':) (loxl) 6EL (X* - X1)2 1 _____EML aEML = (x* - xi) (dX - dXl) + (x* X) (d xl- x*) (l1C) In the section describing the laboratory techniques it was shown that the percentage error in the concentrations, xo and xl differed significantly due to the variation of the error in the analysis with the amount of excess barium hydroxide in the sample. The error in x~ was one percent or less while the possible error in X1 was determined to be as high as six percent. Equation(llC) has been rewritten in the following form dEm - (EM, _ 1) d (dxo - Edx) (12.-' -~I = (Ear - 1) (x - (xox + * ) (lxC) dEML ( l) XlExl (x. x -Ex*) E 1 (xx*X1) Max. % Error in EML E - x) [x x + EMLx*x* ( 1)] ML EML (. -"xI- x _ FxL - X)] (15C) From the results in Table III-C, it is concluded that the maximum error in the plate efficiency, EML, may vary between a minimum of 1.5 and a maximum of 11 percent depentrding on the value of the plate efficiency and the equilibrium liquid concentration. In terms

-248of absolute error, the results in Table III-C indicate that the efficiency data are acc'ucr " vithin +2?-. 5 efficiency percent. TABLE III-C VARIATION OF THE PERCENTAGE ERROR IN PLATE EFFICIENCY, EMLs WITH THE FOLLOWING ERRORS IN x0. x. AND x*. -5 Absolute Error in x and x* = 2 x 10'O 0 x0Absolute Error inx = 4 x 10-5 (X*- x1)x 105 E Error in EML Absolute Error in EML~ 200 o.50 5 2.5 300 0.50 3.3 1.65 400 0.50 2.5 1.25 200 O.75 3.0 2.25 300 0.75 2.0 1.50o 400 0.75 1.5 1.12 200 0.25 11 2.75 3 Oo 0.25 7 4 1.85 400 0.25 5.5 1.37

APPENDJX D CALIBRATION DATA TABLE ID CALI-BRATION DATA FOR ROTAMETER NO. D6-2445 WITH FLOAT NO. D8 21617, STAINLESS STEEL TYPE 347; CALIBRATION BY USING WATER (*) Rotameter Scale cR (***) Reading (**) 42 2.567 57 2.032 64 1.891 72 1.751 94 1.542 L12 1.422 L17. 4o04 L30 1.337 L47 1.273 155 1.245 165 1.220 187 1. 182 186 1.179 202 1.158 219 l 147 232 1.125 264 1. 104 264 1.075 (*) Calibration obtained as follows: The rate of water flowing through the rotameter at each rotameter reading was determined by collecting and weighing the water for a certain timne interval. These data were corrected to voIumetric rate of a gas with a gravity of 0.877 at 14,7 psia and 60~OF by use of the rotameter equation, W = CR, f-p. Data obtained by Dennis Ward(91). (**) Ft3/min. of a gas with a gravity of 0.877 at 14.7 psia and 60~F. (**) Actual flaw rate, ft3/min CR(rotameter scale reading)7 where P1 = density of a gas with a gravity of 0.877 at 14.7 psia and 60~Fo P2 = density of gas being used. -249

-250TABLE II-D CALIBRATION DATA FOR ROTAMETER NOo D6-2445 WITH FLOAT NO. D8-1617, STAINLESS STEEL TYPE 347; CALIBRATION BY ROTARY METER (*) Rotameter Scale Reading (**) CR ( 50 20o814 6o 1 o8307 100 1 3744 100 Lo 3842 150 1.1414 150 1.1612 200 1o 0490 200 1.0595 200 lo 0626 250 1o 0092 250 1o oo76 250 1.0193 300 o 9625 300 0.9800 300 0.o 9900 350 0.9525 350 o09510 350 0.9597 400 0o 9394 400 0o9412 400 0.9347 400 0.9344 () Rotameter calibrated using natural gas and a Roots-Connersville rotary meter; Serial No. 594339 Size, 5 x 15; ft3/revolution, 0.5144; ft3/hr at 1" H20 differential, 23,000; meter located at "J" Station, Michigan Consolidated Gas Co., Detroit, Michigan.r Natural gas gravity, 0.618. (**) Ft3/min of a gas with a gravity of 0.877 at 14.7 psia and 60~Fo (***) Actual flow rate, ft3/min = CR (rotameter scale reading) IPJ —2where Pl = density of a gas with a gravity of 0.877 at 14 psia and 60~F. p2 = density of gas being used.

o CALIBRATED BY USING WATER. CALIBRATED BY ROTARY METER. 2.0 c0 O i LL IO w 1.0 0 40 80 120 160 200 240 280 320 360 ROTAMETER READING Figure 1-D. Calibration of Rotameter D6-2445

-252TABLE III-4D CALIBRATION DATA FOR ROTAMETER W70-4024/1; CALIBRATION BY USING WATER (*) Scale Reading (**) Water Rate, ft3/sec. 20 0.01422 40 0 02845 60 0o 04265 80 0.05693 (*) Date obtained by Ashby(7). Water temperature 54~F; float -density, 7.85 gm/cc. (**) Percent of maxim flw (32.0 gal/in). TABLE IV-D CALIBRATION DATA FOR WET-TEST METEtS, H9SS AND J5SS Meter No. Meter Correction Factor H9SS 0.986 H9gss 0.9835* J5SS 1.0o03 * Value used by Ashby(7)

APPENDIX E PHYSICAL PROPERTIES TABLE I-E VISCOSITY OF CYCLOEXANOL* - Temperatwure, 0C. Viscosity cp 20.4O 70.30 21.00Q 67.10 21.05 66.50 25.80 55.00 25.20 49.20 28.50 42.90 31. 30 35.60 35.80 27.20 38.20 23.82 39. O 22.85 40.15 21.40 42:.86 17.63 42.90 17.58 46.50 14.49 Determined by use of ASTM Meth2od D -253

-25400 0 80 70 60 O 2z w 50 40 30 20. X 10 15 20 25 30 35 40 45 50 TEMPERATURE, ~C Figure 1-E. Viscosity of Cyciohexanol

-255TABLE II -E DEhSITY OF CYCLOHEXANOL* Temperature, "C. Density, gm/cc 46.1 0.9290 47.1 0.9284 46.6 0.9288 45.4 O.9299 42.0 0.9320 41.6 0.9320 41.1 0.9325 40 5 o0.9331 39.7 0o9537 38.9 0.9339 37.7 0.9360 6.1 0.9362 35.9 0.953 35.8 o0,9366 3554 0.9378 34.3 o5 9391 30.8 0.940o3 30.6 09405 30.0 O.9409 29.4 0.9413 28.9 0. 9418 28. 2 O.9421 27.1 0.9431 26.5 0.9435 25.1 0.9443 23.5 0.9459 * Deternmined by ese of a hydrometer, checked by use of a pyCnometer

-256 - 0.950 0.945 PL 0.9637-0.7587t x10-3 0.940 O0 0.935 0.930 I5 20 25 30 35 40 45 50 TEMPERATURE, OC Figure 2-E. Density of Cyclohexanol

-257TABLE III -E SOLUBILITY DATA FOR CARBON DIOXDE IN CYCLOEXANOL (55) Temperature H x10 H x 10-2 ~"~C ~ atm. ft3 atm'lb-:.... mol fraction 25. 1*8~0 2.258 28.3 3.975 2'337 30.8 4.048 2.377 33-.6 4. 181 2.448 36,3 4.362 2.549 39.4 4o475 2,608 41.8 4.624 2.690 47.8 5.-120 2.910 TABLE IV-E VAPOR PRESSURE OF CYCLO'HEXANOL (67) I~ - ~ ~ -~I-I: ~ -; -i i ~:1~ ~ -- i - - --- -:'' ~.,..',T.,~ Temperature ~C Vapor. Pressre, mi Hg 21.0 1 44.0 5 56.0 O10 68.8 20 83.0 40 91.8 60 TABLE V-E VAPOR PRESSURE OF ETHYLENE DIBROMIDE (67) Temperature 0C Vapor Pressure, mm Hg 20 o10.4 25 13.18 30 16.98 35 21.68 40 27.48 -45 34.59 50 43.25

-2583.0 2.9 2.8 2.7 2o.6 2.5 0 2.4 2.3 2.2 2.1 25 30 35 40 45 50 TEMPERATURE,~C Figure 3-E. Henry's Law Constants for Carbon Dioxide-Cyclohexanol System

5. 2 5.0 4.8 4'. -I 4.4 4.2 N~~~~~~~ 4.00 3.8 20 25 30 35 40 45 50 TEMPERATURE,OC Figure 4-E. Henry's Law Constants for Carbon Dioxide-C3/clohexano1 System

APPENDIX F SAMPLE CALCUIATIONS Vaporization Data - Run No. E-10-A 1. Gas Equilibrium Composition, y Pressure above test plate = 305.33 "Hg Avg. liquid temperature on test tray = 33.15~C Vapor pressure of liquid at avg. liquid temperature = 0.780 "Hg O. 780 y* = -0.02572 2, Gas Outlet Composition, Yl Measured sample volume (wet test meter rdg) = 1.000 ft3 Meter correction factor = 0.986 Corrected sample vole = (0.986)(1,000) = 0.986 ft3 The lb moles of dry, inert gas per cubic foot of gas metered were calculated assuming that the gas was.saturated with water vapor at the temperature and pressure of the meter and assumingthe ideal gas law In this:case, the meter temperature was 82:.90F; the meter pressure was 29.08 in. Hg; and the.lb-mole of air per cubic foot of gas metered was 235.77 x 105. Amount.of air in sample (0.986) (235577x: 105 ) = 232.47 x 10-5 lb mole Weight of -ethylene dibromide.condensed in U-tubes 3. 4727 grams = 4.0-748 x 10-5 lb mole Total moles of inert gas plus ethylene dibramide vapor 232,. 47 x 10-5 lb moles = 4.0748 x 10-5 lb moles of C2H4Br2 23565448 x 10-5 lb moles -260

-2614.0748 x 10-5'1 = 236054 x 10-5 = 0 017227 3. Gas Inlet Composition, Yo.. ~.. [ I. I.,. I,' L r, r The method used to deterine the gas inlet composition was identical to that used to determine the outlet composition shown aboveo o, = 0.008166 4. Point Vapor Efficieacy The point and plate efficiencies were assumed to be identical for these runs, i.eo EOG = EMV. Therefore, Yl Y, E, =... - from Equation EOG 1.7227 -816.7 90 60 or 51 25'72.: -56. 7 -755.3.5162 or 51.62% 5. Number of Individual Gas Transfer Units per Plate, NG NG = 1 (1- Y)( o) according to Equation(315:L -y~" (l - Yo)(*. Yl:) 1 1 (1 - 0.017227)(Q.02572 - 0.00866 NG = 1 0.0257 n - io166)0.02572 -'0 0.017227) NCG 0.745 NG = l.n (1 - EoG) according t~o Equatiron(96) NG - In (1l - 0.5162)-= 0.7262 6. Gas Flow Rate The gas rotameter calibration in Table III-D- was applied gases with different Idensities acc:ording to the followingequation:: Q1- = 2 \j pcl

_262where Q1 = volumetric flow rate of fluid 1 at a given meter reading, ft 3/min, Q2 - volumetric flow rate of fluid 2 at the same meter reading,. ft3/min P1 = density of fuid 1, lb/ft3S and P2 = density of fluid 2, lb/ft35o The.actual gas flowing through the r.otameter is nitrogen.-containing a little ethylene dibromide vapor. Its composition is the same as the inlet gas composition, yo. Average molec ularweight of gas flowing through meter yO MC2H4Br2 + (l Y') M2 (o.008166) (1590884) + (1 - 008166) (28) 29.306 Gas rotameter scale reading = 80.5 Temperature of gas in meter = 541.3~R Pressure in.meter -= 31-.08 in Q = 80..5' 6 5= 5.29 ft3/min. The flow rate.may be converted to lb-moles per minute as follows: (75.29) (31.08).V = 0.1961 lb-moles/mine This..is the same as the inlet gas flow rate), Vo, since no vapor is gained or lost.between the rotaneter and the test column inlet.- Therefore Vo = 0.1. 961 lb -moles/mino. The -flow rate of gas above the test; plate, V1- is obtained as follows

-263V1 v0 (1 Yo) = ( = 0.1978 lb moles/min. ~-TCT (0.9528) The average gas flow rate through the froth on the test plate is taken as the arithmetic average of V1 and V.Vavg =.1970 7. Average Gas Temperature The temperature of the gas leaving the tray was calculated with the follQwingrelation: T T - EOG (To Ts) T1 outlet temperature of gas, OF, T = inlet temperature of gas, ~F, and 0 Ts = adiabatic saturation temperature (temperature of liquid on the test plate), F. The inlet gas temperature, To, was 122.0OF T = 122.0 0.5162 (122.0 -91.7) 106.9~F The average gas temperature = 114.4~F or 574.4~R 8. Superficial Velocity If the active plate area is taken as the area between the splash baffle and the inlet doncomer (.615 ft ) and the average pressure is taken azs the pressure above the plate (30,33 in. Hg), V can be converted to gas superficial velocity as follows: avg (oI.Al)(359)(29.92)(574.i) Superficial velocity, vs = (60)(0615)(5 0..3)(..92) = 2.209 ft/sec.

9. Average Gas Densityr, PG Mavg 293 lbs/lb-mole Tavg = 574 460R 28.7 0.33)(4921,5 PG = 300 09 X 155 (-29-=.9)(5I) 00 b709 lbs/ft3 10. F-Factor F = v/ F 2.209V 00709 = 0.5876 11, Gas Contact Time Froth height, Zf =- 4.2 inches Average clear liquid height, Zc - 1.82 inches Gas holdup = Zf Zc 2.38 inches GaS contact time Zf = Zc 238 (12 5 12 2.9 = 0.08978 see. 12. Mass:Transfer Coefficient, kGa NG where tG = gas contact time kGa = 0822 = 8.095

TABLE I -F LABORATORY DATA SHEET - VAPORIZATION DATA Date June 10, 1957 Run No. E-lO-A System N2 - C2H4Br2 Operator Begley - Salesin Weir Height 1-1/2-" Splash baffle height 2" Gas rotameter rdg. 80.5 Liquid rotameter rdg. 40 Barometer 29.08 "Hg at 8090~F Pressure in gas rotameter 2. 00 "Hg Pressure above lst plate 1.25 "Hg Pressure drop across 1st plate T; t "O2 Froth height 420" Clear liquid height, inches (1) -2.00 (2) 1.68 (3) 1.78 (4) 1.80 Temperatures: Position 0C TC mv ~F Liquid on plate No, 1 of rect. col. (in) 33.20 Liquid on plate No. 1-of rect. colo (out) 33.10 Liquid in liquid rotameter No. 1 L-9 1.182 5.1 Gas below plate No. 1 of rect. col. 49-65 L-L 3 1 Gas above plate No. 1 of rect. col. G-2 1.7 Gas in gas rotameter No. 1 Out In Meter No. H9SS J5SS Temperature of meter, ~F - 82_09- T. Pressure in meter, "H20 0,0 0,0.Final meter rdg, cu. ft. 1.o00- 1.000 Initial meter rdg, cu. fto 0.0000 0. 000 Total flow, cuo ft. 1.000: 1.000 Meter factor ____ 1.-03 Drying tube numberT-2 T-12 T-3 S-l2 S-23 S-20:Final tube wt., gms 119.6720 125.4782 144.6585 114.0492 120.4477 115.2178 Initial tube wt., gms'i6.380 125.294 144655 112.5727 120.20642 115.2164 Condensate, gms 5.2910 7i780 0, 0029 14 4765 05 0Sample delivered, cu. ft. 0.2 0.4 0.6 0.8 1.0 Liquid on plate No. 1 (in),~C 5.o20 337c 520 5.20 Liquid on plate No. 1 (out),0C CI 510 33510 55.10 53.10 Gas Temperature, (in), c 50.O00 50.10 5T 50.20 Avg. Liquid temperature, 5.15'C, 91.67 F Avgo Gas temperature (in), 50.05 C 122.09F Inlet sample condensate 1. 6614 grams Outlet sample condensate _.4727 grams

-266TABLE II-F LABORATORY DATA SHEET ABSORPTION DATA Date December 10, 1956 Run # _15A System CO2 CyGelohexanol Operator Begley - Salesin Weir height 3-1/2" Splash baffle height 4" Gas rotometer rdg.o. 41 Liquid rotometer rdgo 82 Barometer 2 "FrHg at 70OO0F Pressure in gas rotomet-er 1o5 Hg.Pressure above 1st plate ~L "Hg Pressure drop across 1 plate lo- 0" H20 Froth height': Clear liquid height, inches ( (2) 5.4 (3) 5_5 (4) Temperatures: ~C Ts m Position -.. Liquid:n plate #1 of recto Colo (in) Liquid on plate #1 of rect, col. (o.ut) 0..951 L7. Liquid entering bottom of recto col.o Liquid in liquid rotometer #2: - Gas below plIte #1 of rect. col. L-3 Gas below plate #1 of:r.ect col (wet bulb) Gas above plate #1 of rect. Col. (dry bUlb) __ G-2: 0.903 _ Gas in gas rotometer #1 1184 Spl. bottle no. 06' 1-05 106A.107 108 Vol. Of BaOH)2,cc. 50 50 50 50 50 50 Splo position. 2 3 5 Vol. of liquid splo -, 100 1 100 O-O. Gas Sample Bulb No.o 2

-267Absorption Data - Run No. 15-A 1. Concentration of Carbon Dioxide in Entering Liquid, X1 Volume of 0.09575N Ba(OH)2 = 50 ml. Volume of 0.09510N HCL = 36.53 ml Milliequivalent of Ba(OH)2 4.788 Milliequivalent of HC1 3.475 Milliequivalent of carbon dioxide 1.313 Syringe volume of liquid sample = 100 cc Correction factor = 0.9957 Actual volume of liquid sample = 99.57 cc Concentration of carbon dioxide 1313 99.57 0.659 x 10-5; moles 41.0 x 10-5 lb. moles 3, lb. moles of cyclohexanol/ft3 = 0.5880 Mol fraction of carbon dioxide = 41.0 x 10-5/0.5876 X1 = 69.7 x 10-5 2. Concentration of Carbona Dioxide in Outlet Liquid Samples, x and x6 2 - Calculation similar to that for x1. x5 = 111.2 x 10-5 x6 = 127.0 x 10-5 3. Equilibrium Concentration of Carbon Dioxide in Liquid Average liquid temperature = —24.00C Pressure above the tray = 1.012 atm.

MPole fractliof o''fcarbon dioxide in gas above the t ray- 0 g9005,Partial. pressure of carbon dioxide above the tray. 0.9113 atm. Henry's Law Constant - 224.0 0 at. mol fraection P 0o9113 Equilibrium mol fraction = H = 2 = 406,8 x l05 40 Murphree Liquid Efficiency,: EML L15 x'* - x 1 111,2 - 69.7 _- 0:]2 — 9o7 = 0.1231 or 12.31% X6 -Xi 127o0 -69.7 17_ 0% EML16 X* - x. 406.8 - 69.7 5. CalculatiOn of Liquid Rate, GPM and LM Ashby(7) used water to calibrate the liquid rotameter, These data are presented in Table, IIID and were used-to calculate the flow rates for cyclohexanol by use of the following equation: Q-l P(PF- P2 where = volumetric flow rate of fluid 1 at,a given meter reading, Q2 = Pvolumetric f.low rate of fluid 2 for the same reading, p1 = density of fluid l, p2 = density of fluid 2 = density of the rotameter floato

_269Rotameter reading = 82.0 Q2 = 0o.0582 (TableIII-D) P1 = 0.945 P2 = 0.99949 PF = 7.85 Liquid rate 0.0582 029949 (7.85 0.9450) 0.9450 (7.65 - 0.99949) = 0.059946 ft3/see = 0.059946 x 7.46 x 60= 26.904 GPM. (o.05995 ),(o.945o)(62.4)(60) 0(.o2o99) ( 0)(64) ( 2.1167 lb-moles/min. LM 100.2 6. Gas Rate, Us, F-Factor, and Gm Gas rotameter reading (Rotameter No. D8-1609) = 41.0 Mol fraction of carbon dioxide = 0.902 M1ol fraction of cyclohexanol in gas (assuming gas to be saturated at plate liquid temperature and pressure) Vapor pressure of cyclohexanol at 24.40C = 0.051" Hg Pressure above plate = 30.28" Hg Mol fraction.cyclohexanol = 0.051 = 0.001684 30.28 1ol fraction air = 1 - 0.902 - 0.001684 0=.096 Molecular weight -of gas = (100.2)(0.001684). + 44 (0.902) + (28.9)(0o096) o.1687 + 39.688 + 2.7744-.=42.631 Pressure at. rotameter = 30068" Hg Temperature at rotameter = 85.10F. -= 1\., = 41 l.48)(545l) = 31.96 ft3/min

-270Gm ( == 0. 08244 lb mole/mia Superficial Gas Velocity Pressure above tray = 30.28&" Hg Temperature above tray = 75o92~F (Gm)(RT) (0 08244)(53509) - =Oo,8638 ft/sec. s - (o.6l5)(60)(P) (1.689)(30628) (P) (M) (30o.28)(42.63 ) 04 b/ft3 pG = (21o82)(T) (2Lo:82)(535.9) F- FaCtor F-Factor = 08638 01104 = 0.2870 7. Murphree Vapor Plate Efficiency, EMV EMV HGIIM 15 EML5 ( 5) - 0.o,1231.1231 + (22o)(o0824)1231) (1. 012) (2. 1167) 0O1231 0.1231 + (8,62)(o08769) 8. Muzphree Vapor Point Efficiency, EOG 0 — x 5 X c2 + c3 + 04 + C5 Xav = (, avg

-271(41.45 + 41.o + 63.5 + 65.5) x 10-5 4 (0o.5885) (52,862) x 10-5 - 89 8 x 10-5 0.5885 x* = 406.8 x 10-5 x5 = 111.2 x 10-5 15 9.8 - 406.8 317.0 1 073 EOG 111.2 - 40o68 295.6 EG1 0.1603 =. o.0193 r 1. 493% EG 1.073.o073 9. Number of Over-All Gas-Phase Mass Transfer Units NOG = - n (1 - EoG) NOG = - n ( - -.01493) NOG 0.0151 1Q. Number of Liquid-Phase Mass Transfer Units NOG NG NL 1 NG NL = NG = (8.62)(0.0151) = 0,130 11. Liquid Contact Time 261.904 GPM Liquid rate - = 2.94x G oP = 0.05991 ft3/see Froth height = 9.0 O1 inches Equivalent cap height = 0.27 inches Relative froth density - - O 6418 Contact time = (9.01 - O. 27)(0.615)(0.648) 4.843 sec. (12)(0.05991)

-2722. Mass Transfer Coefficient NL = kL a tL NL 06 830 tL 4-843

APPENDIX G EXPERIMENTAL AND CALCULATED DATA -273

TABLE I-G PLATE EFFICIENCIES IN RECTANGULAR COUM AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA FOR N2-ETHYLENE DIBROMIDE SYSTEM) 1.50-IN. WEIR, 2.0-IN. SPLASH BAFFLE RUN NO. E-6-A E-6-B E-2-A E-2-B E-1-A E-1-B E-3-A 1. GAS ROT. RDG. 34 36 46 46 77 77 77 2. LIQ. ROT. RDG. 40 40 40 40 38 38 39 3. BAROMETRIC PRESSURE, IN Pg 29.27 29.27 29.32 29.32 29.36 29.36 29.32 4. PRESSURE IN GAS ROT., IN H9 2.10 2.10 1.70 1.80 2.10 1.20 2.05 5. PRESSURE ABOVE PLATE, IN Hg 1.60 1.65 1.20 1.30 1.40 1.10 1.40 6. PRESSURE DROP ACROSS P4ATE, IN 20 3.10 3.10 3.20 3.20 3.60 3.60 3.60 7. FROTH HEIGHT, IN. 3.5 3.5 3.55 3.7 4.2 4.2 4.2 8. CLEAR LIQUID ET. AT POSITION 2, IN 1.90 1.95 1.90 1.95 1.90 1.90 1.95 9. 3 1.80 1.80 1.80 1.80 1.60 1.60 1.65 10. 4 1.82 1.82 1.80 18 1.75.1.80 11. 5 1.80 1 1.81.8 1.80.8 1.75 1.75 1.80 12. AVG. CLEAR LIQ. HT. IN. 1.83 1.83 1.83 1.84 1.75 1.75 1.880 13. AVG. LIQ. TEMP. ON PLATE, IF 93.1 92.98 93.31 93.07 93.02 93.34 92.88 14. LIQ. TEMP. IN ROT. ~F 84.9 85.0 85.3 84.5 80.8 82a4 90.77 15. TEMP. OF GAS BELOW PLATE, ~F 135.3 135.2 131.3 130.9 129.2 130.6 137.0 16. TEMP. IN GAS ROT. ~F 77.0 77.0 78.6 78.0 75.58 76.1 79.0 17. SAMPLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET ILET OUTLET INLET OUTLET INLET 18. TEMP. OF GAS SAMPLE METER, ~F 79.0 78.5 79.5 77.8 77.5 77.3 77.5 77.2 72.4 71.5 73.4 72.5 78.5 79.0 19. PRESSURE IN METER, IN Ho..-.- __ _.. _. -.... — _ __ _ 20. METERED SAMPLE VOL. FT3 1.2000 1.0000 1.101 1.0000 1.0004 1.0000 0.7933 0.8518 1.003 1.0000 0.7765 0.9055 1.0000 1.0000 21. METER CORRECTION FACTOR 0.986 1.003 0.986 1.003 o.986 1.003 0.986 1.003 o.986 1.003 0.985 1.003 o.986 1.003 22. CORRECTED SAMPLE VOLUME FT3 1.183 1.003 1.0856 1.003 0.9864 1.003 0.7822 0.8544 0.9863 1.003 0o.7656 0.9082 -0.986 1.003 23. LB. MOL OF GAS PER FT3 MET. X 105 240.24 240.62 239.89 241.11 241.76 241.90 241.75 241.98 245.69 246.32 245.00 245..60 241.06 240.62 24. LB. MOL OF GAS IN SAMPLE X 105 284.25 241.34 260.42 241.83 238.47 242.63 189.09 206.75 242.32 247.06 187.57 233.05 237.69 241.34 25. ETHIEENE DIBROMIDE ABS.IN TUBES,gm 4.4036 2.3035 4.o436 1.5231 3.8569 1.4533 3.4407 1.7695 4.5938 2.3813 3.6164 2.1744 3.6065 1.5330 26. ETHYLENE DIBROMIDE ABS. IN TUBES, LB. MOL X 105 5.167 2.703 4.745 1.787 4.5256 1.7053 4.0372 2.0763 5.390 2.794 4.243 3.255 4.2307 1.7988 27. TOT. MOLS IN SAMPLE LB.MOL X o105 289.42 244.04 265.17 243.62 243.0 244.33 193.13 208.83 247.71 249.85 191.81 226.31 241.916 243.14 28. Yo0 Yl O.01784 * o.ollo8 0.01789 0.007335 0.01862 o.oo006980 0.02090 o.oo009943 0.o02176 0.01118 0.02212 0.01438 0.01748 0.00738 29. VAPOR PRES.OF LIQ. ON PLATE IN Hg 0.80" 0.80 O.82 0,.805 o.80 0.81 0.80 30. y* 0.02592 0.2587 0.02687 0.02629 -0.02601 0.02659 0.02604 31. E % * 56.58 56.95 58.43 67.03 71.34 59.71 54.12 Am 32. NOG = -n (1-EoG) * 0.855 0.844 0.878 1.11 1.250 0.908 0.779 33. vs, FT/SEC 0.9458 1.o006 1.301 1.284 2.141 2.144 2.157 34. PG LB/F3 0. 07109 o. 07016 0.0694 0. 07106 0. 07197 0.07213 0. 06995 35. F = p 0.2522 0.2668PG 0.3432 0.3422 0.5742 -0.5724 0.5705 56. tG, SEC 0.1471 0.1375 0.1102 0.1210 0.09536 0.09523 0.09272 57. kGa, SEC- * 5,68 6.14 7.97 9.17 13.11 9.535 8.40 * INLET SAMPLE DATA TAKEN FROM E-6-B.

TABLE I-G (CONTINUED) -PLATE EFFICIENCIES IN RECTANGULAR COLL94 AT MUNIESITY ~'MICHIGAN (ORIGINAL DoaT FOR X2-ETmYLENE DIBROMIDE SYSTEM) 1.50-IN. WEIR, 2-.O-IN. SPLASH BAFFLE RUN NO. F _ -3-B E-4-A E-J~-B E-5-A E-5-B E-11-A E-11-B 1. GAS ROT~. REG. 77 114 114. 161 161 346.5 46.5 2. LIQ. ROT. REG. 56 39 40 40 40 4O0 4O 3.- BAROMETRIC PRESSURIE, IN Fg 29.32 29.30 29.30 29.15 29.15 28.87 28.87 4. PRESSURE IN GAS ROT., IN Hg 2.20 2.20 2.05 2.60 2.-60 1.45 1.60 5. PRESSURE ABOVE PLATE, IN Ag 1.45 1.22 1.10 1.05 t. -05 0.90 1.00 6. PRESSURE DROP ACROSS PLATE, IN H20 3.60 4.48 4.33 6.10o 6.10o 3.15 3.15 7. FROTHi HEIGHT, IN. 5.9 5.2 5.2 6.5 6.5 3.7 3.7 8.CLEAR LIQ. HY. AT POSITION 2, IN 1.90 2.0 2.00 2.05 2.05 2.00 2.00 9. 3 1.60 1.5 1.50 1.4-5 1.40 1.83.1.83 lO. 4 -1.75 1.7 1;7o 1.4o 1.4o 1,-83 t.85 ll. 5 1.75 1.8 1.80 1._70 1.70 1.80 1.80 12. AVG. CLEAR LIQ. HT. IN. 1.75 1.75 1.75 1.65 1i.64 1.86 18 13. AVG. LIQ. TEMP. ON PLATE, OF 93.02 93.81 94.25 94.7 94.57 92.8 91.987 14. LIQ. TEMP. IN ROT. ~F 9 0. 7 91 ~ ~ ~ ~~ ~ ~ ~ ~~~~~~~~~~~~~~.73 92.9 029. 6 586. 15. TEMP. OF GAS BELOW PLATE, ~F 137.0 134.3 136.2 139.1 t138.4 128. 9 119.4I' 16. TEMP. IN GAS ROT. OF.78.8 78.0O 78.1 76.5 76. 5 82.5 82.9 - - 17. SAMPLING PoINT OUTLET INLET 0 UTLET INLET OUTLET INLET OUTLET INWET OUTLET INUET OUTLET INLET OUTLET ZLTkJ 18. TEMP. OF GAS SAMPLE METER, ~F 79.5 79.0 79.2 78.7 80.3 79.9 78.0 77.6 78.2 80.3 80.1 80.5 80-.7 8. %9. PESSURE IN;NETER, IN %2.............. 20. METERED SAMPLE VOL. FT3 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1,0000 1.002 1.000 1.000.0 21. METER CORRECTION FACTOR O. 986 1.003 O. 986 1.003 O. 986 1.003 0. 98 1.003 0. 986 1.003 O. 986 1.003 0. 9861.0 22.:CORRECTED SAMPLE V0LM~, FT3 0.986 1.003 0.986 1.003 0.986 1.003 0.986 1 003 0.986 1.003 0.988 1.003 0.9861.0 23. LB. M0L OF GAS PER FT3 MET.X 105 240.32 240.84 240.36 240.70 239.56 2-39.86 239.93 2)L0.24 239.80 240.O8 235.92 236.06 235.61 257 24. LB. M0L oF GAS IN SAMPL X 105 236.96 241.56 236.99 241.42 236..21 240.58 236.57 240.96 236.44 240.80 233.10 236.77 233.O1 264 25. ETHYLENE DIBROMIDE ABS. IN TUBES.,gin 3.6502 1.5202 3.7102 1.4884 3.T'{9 1.4594 3.8830 1.6178 3.8521 1.6851 3.7141 1.655 3.7667 1 70 26. ETHYLENE DIBROMIDE ABS. IN TUBES,53 LB. M0L X 105 4.2831 1.7838 4.3535 1.7465 4.4329 1.7512 4.5562 1.8983 4.5200 1.9773 4.3580 1.9396 4.4198 200 27. TOT.MOLS IN SAMPLE. LB.MOL X 105 2.40.24 243.34 241.34 243,17 240.64 242.33 241. 2126 242.86 240. 96 242.78 237.46 238.71 237.4328. 28. Yo, Yt. 1 7 0.007330 0.01804 0. 007182 0.01842 0.007227 0. 01890 0. 007817 0.01876 O. 008145 0. 018353 0.oo8_125 O.o018615 0081 29. VAPOR IRES.0OF LIQ. ON PLATE IN Fg 0.80 0.825 0.830 0.84o 0.835 0.810 O_. 810 30.- y* o.o0260o0 o. 2 7-03 0. 02730 0. 02782 o.o02765 o. 027209 0.027118 31. E0 55.84 54.70 55.75 55.40 54.425.95.0 32. NMG =- In (1-EoG) 0.819 0.886 0. 815 0.807 0.786 Q.7678 0.783 533 vs, FT/SEC 2.1676 3.227 3.206 4.638 4.665. 1.2896 1.'288 34 pG LB/FT3 0.06963 0.0695 0.06923 0.o6865 O.06881 0.06929g 0.07022 35. F = vs'rPG -0-5722 o.85~~ ~ ~~~~~~02. 848 1.2150 1.2222 0.3410o 0.3418 36. t:G SEC -o.O8266 0.09186 0.08968 0.08804 0.088. la.1184 37. k'Ga, SEC-' 9.908 9.645 9.088 9.17 9.053 6.4929.1 *INLET SAMPLE DATA TAKEN FROM E-6-B.

TABLE Z-G (CONTINUED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA FOR N2-ETHYLENE DIBROMIDE SYSTEM) 1.50-IN. WEIR, 2.0-IN~ SPLASH BAFFLE RUN NO. E-7-A E-7-B E-lO-A E-10-B E-9-A E-9-B E-8-A E-8-B 1. GAS ROT. RDG. 78 78 80.5 80.7 119 ll9 158 158 2. LIQ. ROT. RDG. 59.4 59.9 40 40 59 40 ~0 59 5. BAROMETRIC PRESSURE, IN Hg 28492 28.92 29.08 29.08 29.O9 29.09 29.15 29.13 4. PRESSURE IN GAS ROT., IN Hg 5.70 3.40 2.00 2.05 2.25 2.25 2.65 2.65 5. PRESSURE ABOVE PLATE, IN Hg 1.50 1.50 1.25 1.30 L25 1.25 1.20 1.20 6. PRESSURE DROP ACROSS PLATE, IN H20 5.50 5.25 3.60 5.60 4.45 4.45 5.77 5.77 7. FROTH HEIGHT, IN. 4.0 4.0 4.2 4.2 5.5 5.3 6.2 6.4 8. CLEAR LIQ. HT. AT POSITION 2, IN 2.00 2.00 2.00 2.00 2.05 2.02 2.10 2.05 9. 5 1.72 1.75 1.68 1.68 1.50 1.50 i. 40 1.40 10. 4 1.80 1.80 1.78 g 80 1.70 1.70 1.45 1.45 11. 5 t. 80 1.80 1.80 1.80 1.78 1.80 t. 75 1.75 12. AVG. CLEAR LIQ. HT]tN. 1.85 1.84 1.82 1.82 1.76 1.76 1.75 1.66 15. AVG. LIQ. TEMP. ON PLATE, ~F 87.44 87.55 91.67 91~49 92.48 92.59 94~06 94.06 14. LIQ. TEMP. IN ROT. ~F 86.2 85.9 85.1 85~0 85.8 85.9 91.1 91.5 J 15. TEMP. OFGAS BELOW PLATE~ ~F 121.3 120.9 114.6 114.6 114.4 114.7 116.2 116.2 ~ 16. TEMP. IN GAS ROT. ~F 151.5 151.3 81.5 81.1 79.2 79.5 76.5 76.7 -'-,] 17. SAMPLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET O'h 18. TEMP. OF GAS SAMPLE METER, ~F 85.50 85.30 85.10 85.0 82.9 82.5 82.7 82.5 81.1 80.8 81.9 81.4 74.7 81.4 75.0 74.5! 19. PRESSURE IN METER, IN 520............................ -.... 20. METERED SAMPLE VOL. FT5 1.200 1.200 1.2111 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.0002 1.OOOO 0.9998 1.0000 21. METER CORRECTION FACTOR 0.986 1.003 0.986 1.005 O.986 1.005 0.986 1.003 0.986 1.005 O.986 1.003 0.986 1.003 0.986 1.005 22. CORRECTED SAMPLE VOIIIME, FT3 1.i832 1.2036 1.1941 1.003 O.986 1.O05 0.986 1.003 O.986 1.005 O.986 1.003 O.9862 1.003 00.9918 1.005 25. LB. MOL OF GAS PER FT5 MET.X 105 253.98 234.15 254.27 234.35 255.77 236.06 235.95 256.06 256.18 257.45 256.60 256.96 242.14 ~ 242.6 -~4i.94 242.39 24. LB. MOL OF GAS IN SAMPLE X 105 276.85 281.80 279.74 255.05 252.47 236.77 252.65 236.77 232.87 258.16 233.29 257.67 258.80 245.55 241.89 245.12 25. ETHYLENE DIBROMIDE ABS. IN TUBES, gm 3.6217 2.3206 5.6605 1.6900 3.4727~ 1.6614 5.4409!.6769 3.5071 1.6493 5.6396 1.6204 5.7795 1.8531 3.8491 1.9707 26. ETHYLENE DIBROMIDE ABS. IN TUBES, LB. MOL X 105 4.250 2.725 4.295 1.985 4.0748 1.9495 4.0375 1.9676 4.1152 1.9355 4.2706 1.9015 4.4348 2.1744 4.5164 2.3124 27. TOT. MOLS IN SAMPLE LB.MOL. X 105 281.10 279.08 284.04 257.05 236.54 258.72 256.67 2'58.74 256.99 240.09 237.56 259.57 245.24 245.50 _~46.41 245.43 28. Yo' Yl 0.O15119 O. 009757 O. 015121 O. 008566 O. 017227 O. 008167 O. 017556 O. 008242 0.017364 O. 008060 O. 017977 O. 007936 O. 018233 O. 008857 O. 018329 O. 009422 29. VAPOR PRES.OF LIQ.ON PLATE IN Hg 0.699 0.695 0.780 0.78 0.800 0.800 0.835 0.835 50. Y* 0.02298 O. 02500 O. 0'2572 0.D25675 O, 026568 O. 0~6368 O. 02753 0.027531 31. EOG{ 40.55 46.0~ 5~ 62 52.28 50.82 54.48 50.21 49.19 52. N0G = -~n (1-EoG) 0.520 0.6166 0.7262 0.7402 0.7095 0.7974 0.7324 0,677 35. vs, FT/SEC 2.064 2.0712 2.209 2.211 5.290 3.504 4.387 4.400 34. PG LB/FT5 O. 07142 O. 07064 0.07092 0.O7109 O. 07085 O. 07090 O. 07156 O. 07156 35. F = VSWPG 0..5490 0.5509 0.5876 0.5897 O.8751 0.8787 1.1735 1.177...... 56. tG, SEC 0.0876 0.O8691 0.08976 0.09347 0.~D8967 0.08930 0.08491 0.08977 57. k'Ga, SEC'I 5.936 7.095 8.089 7.55 7.921 8.929 8.20 7.54 INLET SAMPLE DATA TAKEN FROM E-6-B.

TABLE i I - (CONTINUED) PLATE EFFICIENCIES TN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA FOR N2-ETHYLENE DIBROMIDE SYSTEM) 5.50-IN. WEIR, L4-IN. SPLASH BAFFLE RUN NO. E-15-A E-15-B E-12-A E —2-B E-17-A E-17-B 1. GAS ROT. RDG. 41 41 45 45 61 62 2. LIQ. ROT. RDG. 40 40 40 40 44 39 3. BABOMETRIC PRESSURE, IN Hg 29.20 29.20 28.87 28.87 29.16 29.16 4. PRESSURE IN GAS ROT., IN Hg 1.56 2.20 1.50 1.85 2.10 2.15 5. PRESSURE ABOVE PLATE, IN Hg 1.00 1.50 0.70 1.05 1.25 1.25 6. PRESSURE DROP ACROSS PLATE,IN i20 6.60 6.45 6.35 6.30 6.15 6.13 7. FROTH HEIGHT, TN. 6.9 6.9 6.7 6.6 6.8 6.9 8. CLEAR LIQ. HT. AT POSITION 2,IN 4.00 3.90 3.68 3.73 3.60 5.60 9. 3 3.55 5.50 5.25 5.28 5.10 5.05 10. 4 5.55 5.50 3.30 3.35 5.20 3.15 11. 5 3.65 3.60 3.38 3.40 5.50 5.25 12.'AVG. CLEAR LTQ. HT. IN. 3.69 3.63 3.40 3.44 3.30 5.26 13. AVG. LIQ. TEMP. ON PLATE,'F 95.88 95.92 92.9 95.0 95.72 94.51 14. LIQ. TEMP. IN ROT. OF 86.5 86.1o 83.4 83.8 88.9 89.5 15. TEMP. OF GAS BELOW PLATE,'F 152.9 133.2 130.8 151.0 151.4 152.0 16. TEMP. IN GAS ROT. OF 91.9 92.0 82.0 85.0 87.4 88.0 17. SAMPLING POINT OUTLET INLET )UTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET 18. TEMP. OF GAS SAMPLE METER,'F 90.5 90.0?1.1 90.5 79.5 79.5 80.4 80.0 85.2 85.0 86.5 86.4 19. PRESSURE IN METER, IN H20 20. METERED SAMPLE VOL. FTS 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 0.9998 0.9998 21. METER CORRECTION FACTOR 0.986 1.005 0.986 1.005 0.986 1.005 0.986 1.003 0.986 1.003 0.986 1.005 22. CORRECTED SAMPLE VOLUME, FT5 0.,986 1.005 0.986 1.003 o.986 1.005 0.986 1.005 0.986 1.005 0.9858 1.0027 23. lB. MOL OF GAS PER FT3 MET.X 105 250.93 250.6 250.4 251.15 256.49 236.49 255.84 256.15 224.7 254.9 233.9 255.9 24. LB. MOL OF GAS IN SAMPLE X io5 227.70 251.29 227.0 252.0 233.18 257.20 252.54 256.84 221.5 255.6 250.5 254.5 25. ETHYLENE DIBROMIDE ABS. IN TUBES gm 4.0816 2.0864 4.1365 1.9519 5.8956 1.7640 4.0157 1.7317 4.1086 2.1495 4.1915 2.0557 26. ETHYLENE DIBROMIDE ABS. IN TUBES, LB. MOL X io5 4.7895 2.4481 4.850 2.27 4.5687 2.0698 4.7096 2.0519 4.829 2.517 4.91 2.41 27. TOT.MOLS IN SAMPLE LB.MOL X 105 252.49 233.74 251.85 254.5 257.75 259.27 257.25 258.87 226.5 258.1 255.41 256.9 28. y0' Y1 0.02060 0.010475 0.02095 o.oog06 0.019217 0.008651 0.019851 0.008506 0.021338 0.01057 0.020858 0.010175 29. VAPOR PRES.OF LIQ.ON PLATE IN Hg 0.830 0.830 0.817 0.82 0.855 0.85 30. y* 0.027458 0.0272 0.27513 o.o27406 o.027458 0.02795 51. EOG% 59.83 64.51 56.02 60.02 65.76 6o.lo 52. NOG = -2n (1-EOG) 0.914 1.0503 0.821 0.917 1.0143 0.92 55. vs, FT/SEC 1.115 1.125 1.295 1.260 1.669 1.681 34. pG LB/FT3 o.o708 0.0705 0.0692 0.06925 0.07139 0.07109 55. F = vsPG 0.289 0.299 0.3419 0.3325 0.4460 0.4480 56. tG, SEC 0.241 0.250 0.2122 0.2090 0.17475 0.1804 37. k'GaSEC-l 3.79 4.121 5.869 4.390 5.958 5.099 INLET SAMPLE DATA TAKEN FROM E-6-B.

TABLE 1-G (CONTINUED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA FOR N,, —ETEYLENE DIBROMIDE SYSTEM, 3.50-IN. wEiR, 4-IN. SPIASH BAFFLE RUN NO. E-13-A E-13-B E-14-A E-14-B E-16-A E-16-B 1. GAS ROT. RDG. 79 79 116 116 144 143 2. LIQ. ROT. RDG. -40 40 40 41 4o 4 3. BAROMETRIC PRESSURE, IN Hg 28.84 28.84 28.92 28. 91 29.120 29.12 4. PRESSURE IN GAS ROT., IN Hg 2.55 2.45 2.20 2.20 2.60 2.80 5. PRESSURE ABOVE PLATE, IN Hg 1.55 1.45 0.93 0. 93 0.90 1.00 6. PRESSURE DROP ACROSS PLATE,IN H20 7.00 7.00 8.20 8.05 10.00 9.60 7.- FROTH HEIGHT, IN. 8.0 8.0 10.0 10.0 11.65 11.7 8. CLEAN LIQ. HT.AT POSITION 2,IN 3.81 3.82 4.00 4.00o 4.15 3,95 9. 3 3.33 3.30 3.65 3.30 3.35 3.15 10. ~~ ~~~~~4 3.37 3.35 3.30 3.30 3.20 3.00 11. ~~ ~~~~~5 3,.50 3.'50 3.60 3.55 3-80 3,45 12. AVG. CLEAR LIQ. HT. IN. 3.50 3.49 3.46 3.54 3.63 3. 59 13. AVG. LIQ. TENT'. ON PLATE, OF 93.2 93.1 93.38 93.42 93.2 92.13 14. LI@Q. TEMP. IN ROT. OF 84.3 83.3 87.2 87.7 84.7 85.0 15. TEMP. OF GAS BELOW PLATE, OF 131.0 131,0 131.7 132.0 129.0 132.0 16. TEMP. IN GAS ROT. OF 82.7 83.0 81.8 82.0 84. 84.2 17. SAMLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET 18. TEMP. OF GAS SAMPLIE METER, 0F 83.0 82.5 83.5 83.2 84.5 84.2 85.5 84.8 91.7 91.5 91.15 9 19. PRESSURE IN METER, IN H20 -- -- -- -- - - - - - - -- -- - - - - 20. MTERED SAMPLE VOL. FT3 1.0000 -1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000.00 21. MTER CORRECTION FACTOR 0.986 1.003 0.986 1.003 0.986 1.003 0. 986 1.003 0.986 1.003 0. 986 100 22. CORRECT-ED SAMPLE VOLUME, FT3 0.986 1.003 0.986 1.003 0.986 1.003 0. 986 1.003 0.986 1.003 0.986 1.3 23. LB. MOL OF GAS PER FT3. MET. X icS 233.67 234.04 233.31 233.5lt 233.25 233.47 232.41 232.93 229.4 229.6 229.5 200 24. LB. MOL OF GAS IN SAMPLE X 105 230.40 234.75 230.04 234.24 229.93 234.17 229.15 233.63 226. 18 230.3 226.2 2 25. ETRffELE DIBROMIDEE ASS. IN TUBEs,gm 4.0598 1.7750 4:0253 1.7738 4.0933 1.7824 4.1316 1.7961 4.0104 1.9660 3.9965 57 26. ETHYLEN DIBROMIDE ASS. IN TUBES, LB. MOL X io5 4. 7637 2.0593 4.7232 2.0802 4.8030 2.0o914 4.8479 2.1075 4.705 2.306 4.690 5,3 27. TOT.MOLS IN SAMPLE LB.MOL X 105 235.16 236.81 234.77 236.32 234.78 236.26 234.00 235.74 230.88 232.6 230.89 23. 28. 0.020257 0.008691 0.020119 0.008802 0.020457 0.008852 0.020718 0.008940 0.020378 0.009914 0.02032.00 29. VAPOR PRES.OF LIQ.ON PLATE IN Hg9 0.82 0.817 0. 82 0. 823 0.82 0. 82 30. y* 0.o26983 0.027017 0.027471 0.027663 0.027315 0.02731 31. E0% 63.23 62.13 62.3-3 62.90 60.13 59.55. 32. NW = -Agn (1-Em 1.037 1.031 1.0125 1.028 0.92 0.906 33. VS, FT/SEC -2.20 2.20 3.250 3.255 4.036 4.3608 -34. pG L/FT 0.070-5 0.0703 0.0694 0.0696 0.0-704 0.07035 35. F = V~4/G 0.585 0.585 0.858 0.862 1.0655 1.155 36. tG, SEC 0.1705.0.1704 o.i168 0:1655 0.1656 0.1588 37. k' Ga, SEC.7 6.080 6.070 6.025 6.210 5.55-5 5.705 *INLET SAMPLE DATA-TAKEN FROM E-6-B.

TABLE I-a (CONTINUED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA FOR N2-ETHYLENE DIBROMIDE SYSTEM) 3.50-IN. WEIR, 4-IN. SPLASH BAFFLE RUN NO. E-18-A E-18-B E-19-A E-19-B E-20-A E-20-B 1. GAS ROT. RDG. 60 60 81 82 115 116 2. LIQ. ROT. RDG. 80 80 79.5 79 79.5 79.5 3. BAROMETRIC PRESSURe, IN Hg 29.10 29.10 29.03 29.03 29.00 29.00 4. PRESSURE IN GAS ROT.,IN Hg 2.60 2.60 2.50 2.-60 3 40 3.00 5. PRESSURE ABOVE PLATE, IN Hg 1.50 1.55 1.30 1.30 1.70 1.60 6. PRESSURE DROP ACROSS PLATE,IN H20 8.25 6.90 9.10 8.80 11.00 10.10 7. FROTH HEIGHT, IN. 9.0 8.5 10.1 10.1 11.4 11.4 8. CLEAR LIQ. HT. AT POSITION 2, IN 4.60 4.5 4.90 4.70 5.50 4.90 9. 3 4.00 3.95 4.10 4.40 4.40 4.05 10. 4 4.10 4.05 4.20 4.10 O 4.70 4.20 11. 5 4.25 4.10 4.30 4.20 4.90 4.30 12. AVG. CLEAR LIQ. HT. IN. 4.i 4 4.15 4.37 4.35 4.88 4.425 13. AVG. LIQ. TEMP. ON PLATE, *F 93.56 93.88 93.60 94.21 93.0 92.8 14. LIQ. TEMP. IN ROT. ~F 86.0 87.4 91.80 93.0 90.4 90.4 15. TEMP. OF GAS BELOW PLATE, ~F 130.8 132-.5 131.1 132.5 130.8 130.6 16. TEMP. IN GAS ROT. ~F 82.4 82.4 81.5 81.5 81.8 81.68 17. SAMPLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET IL 17. SAMPLINGP~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~INET! 18. TEMP. OF GAS SAMPLE METER, ~F 80 79.75 80.2 80.0 80.0 79.8 79.5 79.5 82.2 82.0 82.6 82.3 19. PRESSURE IN METER, IN o0 __ _ _ _ __ __ _ __ __ _ 20. METERED SAMPLE VOL. FT3 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 21. METER CORRECTION FACTOR o.986 1.003 o.986 1.003 o.986 1.003 o. 986 1.003 o.986 1.003 o.986 1.003 22. CORRECTED SAMPLE VOLUME, FT3 0.986 1.003 0.986 1.003 0.986 1.003 0.986 1.003 0.986 1.003 0.986 1.003 23. LB. MOL OF GAS PER FT3 MET.X 105 238.1 238.3 238.1 238.1 237.6 237.6 237.9 237.9 235.7 235.7 235.4 235.40 24. LB. MOL OF GAS IN SAMPLE X 105 234.7 239.0 234.7 238.8 234.2 238.2 234.5 237.9 232.4 235.7 232.1 236.11 25. ETHYLENE DIBROMIDE ABS.IN TUBES,gmn 4.0602 1.9829 4.1501 2.2425 4.1460 1.9727 4.1936 1.9579 4.0114 2.0009 3.9966 1.9404 26. ETHYLENE DIBROMIDE ABS. IN TUBES, LB. MOL X 105 4.76 2.326 4.87 2.63 4.86 2.31 4.40 2.29 4.706 2.340 4.689 2.276 27. TOT.MOLS IN SAMPLE LB.MOL X 105 239.46 241.33 239.57 241.43 239.06 240.51 239.40 240.19 237.10 238.04 236.79 236.10 28. yo, Yl 0.019870 0.09639 0.020328 0.010893 0.020329 0.0096 0.020467 0.009537 0.019848 0.00983 0.01980 0.009639 29. VAPOR PRES.OF LIQ.ON PLATE IN Hg 0.823 0.824 0.82 0.835 O.81 0.81 30. y* 0.026895 O. 026884 O. 027035 O. 027530 O. 0263 O. 026470 31. EOG% 59.46 59.0 61.53 60.74 -6o.8e 60.37 32. NOG -An (1-EOG) 0.899 0.8878 0.9795 0.9315 0.933 0.9269 33. VS, FT/SEC -1.2107 1.2076 2.2509 2.2826 3.192 3.214 34. PG LB/F-T3 O0.07218 0.07259 O. 07084 0. 07069 O.07181 0.07145 35. F = Vs4 PG 0.3245 0.3254 0.598 0.607 0.855 0.8597 36. tG, SEC O. 3278 O. 3001 0.2119 0.2135 0.1703 0 1747 37. k'Ga,SEC-1 2.7425 2.958 4.613 4.362 5.4785 5.305 * INLET SAMPLE DATA TAKEN FROM E-6-B.

TABLE I-G (CONTINUED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA FOR N2-CYCLOHEXANOL SYSTEM) 1.50 IN-. WEIR, 4-IN. SPLASH BAFFLE RUN NO. H —B H-2-'A H-3-A H-3-B 1. GAS ROT. RDG. 120 120 80 80 2. LIQ. ROT. RDG, 25 25 25 25 3. BAROMETRIC PRESSURE, IN Hg 29.23 29.14 29.23 29.23 4. PRESSURE IN GAS ROT.,IN Hg 2.55 2.65 2.20 2.25 5. PRESSURE ABOVE PLATE, IN Hg 1.80 2. 05 1.87 1.90 6. PRESSURE DROP ACROSS PLATE, IN H2O -- - 2.00 2.10 7. FROTH HEIGHT, IN. 5.3 5.3 4.10 4.2 8. CLEAR LIQUID HT. AT POSITION 2,IN 1.95 1.95 2.00 2.00 9. 3 1.535 1.35 1.50 1.55 10. 4 1.35 1.55 1.60 1.60 l1. 5 1.70 1.70 1.70 1.70 12. AVG. CLEAR LIQ. HT. IN. 1.59 1.59 1.70 t1.71 13. AVG. LIQ.. TEMP. ON PLATE, ~F 114.4 114.0 115.9 115.7 14. LIQ. TEMP. IN ROT., ~F 113.1 109.1 113.2 102-9 O9 15. TEMP. OF GAS BELOW PLATE, ~F 114.0 109.8 117.4 115.9 o 16. TEMP. IN GAS ROT.,OF 77.6 75.4 81.1 81.1' 17. SAMPLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET 18. TEMP. OF GAS SAMPLE METER, ~F 78.50 79.00 78.6 78.8 81.5 81.6 81.5 82.0 19. PRESSURE IN METER, IN 20 -- -- -.. -_ _ _ 20. METERED SAMPLE VOL. FT3 -0.9999 0.9997 0.6001 0.7365 0.0231 1.0000 1.0000 1.0008 21. METER CORRECTION FACTOR. 986 1.003 1.003 O.986 1.003 0.986 1.003 0.986 22. CORRECTED SAMPLE VOLUME, FT3 o.986 1.003 0.6019 0.7262 0.6250 0.986 1.003 0.9868 23. LB. MOL. OF GAS PER FT3 MET.X 105 240.1 239.7 239.6 239.8 238.3 238.2 238.3 237.6 24. LB. MOL. OF GAS IN SAMPLE X 105 236.8 240.4 144.0 174.0 149.0 235.0 239.0 234.0 25. CYCLOHEXANOL ABSORBED IN TUBES, gm 0.5708 0.2922 0.3642 0.2035 0.3750 0.3191 0.7149 0.3174 26. CYCLOHEXANOL ABS. IN TUBES, LB. MOL. X 105 1.240 0.644 0.803 0.450 0.825 0.702 1.351 0.699 27. TOT. MOL. IN SAMPLE, LB.MOL.X 105 238.04 241.044 144.80 174.45 149.825 235.7 240.351 234.7 28. Yo, Yl1 MOLS. CYCLOHEXANOL/MOL.GAS 0.00521 0.00267 0.00555 0.00258 0.-00351 0.00298 0.00562 0.002979 29. VAPOR PRES. OF LIQ.ON PLATE, IN Hg 0.2165 0.204 0.2244 0.2205 30. y* 0.00697 0.00681 0.00721 0.00705 31. EOG% 35.8 70.3 59.9 65.0 32. NOG = -Qn (1-EOG) 0.445 1.225 0.922 1.061 33. vs, FT/SEC 3.352 3.32 2.21 2.18 34. pG LB/FT3 0.0701 0. 0702 0. 0698 0.0707 35-. F = vsTPG 0.884 0.882 0.584 0.580 36. tG, SEC o. o928. 0932 0.0999 0.0951 37. k'Ga, SEC-1 4.80 13.14 9.24 11.17 G 1.17

TABLE I-G (CONTI1ND) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA FOR N2-CYCLOBEXANOL SYSTEM) 1.50-IN. WEIR, 2 _IN. SPLASH BAFFLE RUN NO. H-4-A H-4-B H-5-A H-5-B H-6-A E-7-A 1. GAS ROT. RDG. 120 120 10O 120 160 80 2. LIQ. ROT. RDG. 25 25 25 25 25 25 3. BAROMETRIC PRESSURE, IN Hg ABS. 29.3-0 29.30 29.40 29.4 29.5 29.065 4. PESSURE IN GAS ROT.,IN Hg 1.85 2.00 2.45 2.80 3.30 2.00 5. PRESSURE ABOVE PLATE, IN Hg 2.50 2.60 1.85 2.00 2.20 1.60 6. PRESSURE DROP ACROSS PLATE, IN H20 3.00 2.80 2.90 2.80 3.90 2.00 7. FROTH HEIGt, IN. 5.50 5.5 5.7 5.7 7.0 4.4 8. CLEAR LIQ. SrT., POSITION 2, IN. 1.90 1.85 2.10 2.10 2.15 2.15 9. 3 1.50 1.55 1.55 1.55 1.50 1.70 10. 4 1.50 1.55 1.55 1.55 1.50 1.80 11. 5 1.90 1.90 2.50 2.30 2.10 1.90 12. AVG. CLEAR LIQ. HT. IN. 1.70 1.71 1.81 1.81 1.81 1.89 13. AVG. LIQ. TEMP. ON PLATE,'F 101.35 101.44 100.76 100.78 99.77 100.653 14. LIQ. TEMP. IN ROT. ~F 95.5 95.4 97,7 97.9 95.5 100.0 O 15. TEMP. OF GAS BELOW PLATE, ~F 108.0 108.5 106.5 107.0 105.0 106.0 16. TEMP. IN GAS ROT.,~F 78.o 78.1 76.6 76.5 95.4 77.1 17. SAMPLING POINT OUTLET INIET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET 18. TEMP. OF GAS SAMPLE METER, ~F 83.8 85.2 83.8 84.2 78.2 78.2 78.4 78.4 71.0 71.0 72.5 72.5 19. PRESSURE IN METER IN 20 -- -- -- -- -- -- -- -- 20. METERED SAMPLE VOL.FT3 1.000 1.000 1.0325 1.0000 1.00 1.000 1.0000 1.0000 1.000 1.0000 1.0000.00 21. METER CORRECTION FACTOR 1.003 0.986 1.005 0.986 1.003 0.986 1.003 O.986 0.986 1.003 0.986 1.003 22. CORRECT SAMPLE VOL. FT3 1.003 0.986 1.038 0.986 1.003 0.986 1.003 0.986 0.986 1.003 0.986 1.003 23. LB.MOL.OF GAS/FT3 MET.X 105 237.15 237.75 257.15 236.65 241.6 241.6 241.6 241.6 243.3 245.8 243.0 243.0 24. LB.MOL.OF GAS IN SAMPLE X 105 238.0 234.0 246.0 234.0 242.0 238.0 242.0 238.0 239.9 244 240.0 244.0 25. CYCLOHEXANIOL ABSD.IN TUBE, gm 0.5057 0.3174 0.4618 0.3147 0o4325 0.2976 0.4335 0.2925 0.4205 0.2402 0.4932 0.3011 26. CYCLOHEXANOL ABSD.IN TUBE, LB. MOL. X 105 1.115 0.699 1.015 0.694 0.952 0.655 0.954 0.645 0.925 0.53 0.889 0.663 27. TOT.MOLS.IN SAMPLE LB.MOL.X 105 239.115 234.699 247.015 234.7 242.4 238.3 242.4 238.65 240.83 244.5 240.9 244.7 28. yo, Y1 MOLS.CYCLOHEXANOL/MOL.GAS 0.01678 0.00298 0.00412 0.002952 0.00392 0.00274 0.00392 0.00270 0.00384 0.oo02168 0.00369 0.00271 29. VAPOR PRES. OF LIQ.ON PLATE,IN Hg 0.138 0.138 0.134 0.135 O.2 -8 0.132 30. y* 0.00434 0.00oo433 0.00428 0.00430 0.00404 0o. oo427 31. EOG% 124. 1 84.8 76.7 76.3 89.3 62..8 32. NOG = -n (1-EOG) -- 1.882 1.436 1.440 2.24 0.995 33. vs., FT/SEC -- -- 3.28 3.288 4.19 2.165 34. pG LB/F.P3 o 0.0716 0.0716 0.0725 o.0708 35. F = VsP'G -. __ 0.877 0.878 1.130 0.576 36. tG, SEC -- -- 0.0989 0. o0986 0.103 0.0966 37. k'Ga, SEC-1 -- 14.52 14.60 21.5 10.3

TABLE I-G (CONTnuED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA FOR N2-CYCLOHEXANOL SYSTEM) 1.50-IN. WEIR, 2-IN. SPLASH BAFFLE RUN NO. H-7-B H-8-A H-8-A H-9-A H-9-B 1. GAS ROT. RDG. 80 40 40 160 160 2. LIQ. ROT. RDG. 25 25 25 25 25 5. BAROMETRIC PRESSURE, IN Hg ABS. 29.065 29.07 29.07 28.86 28.86 4. PRESSURE IN GAS ROT.,IN Hg 2.20 2.10 2.00 2.65 2.65 5. PRESSURE ABOVE PLATE, IN Hg 1.85 1.80 1.80 1.60 1.60 6. PRESSURE DROP ACROSS PLATE, IN 1120 2.10 1.60 1.55 4.00 4.oo00 7. FROTH HEIGHT, IN. 4.35 3.5 3.5 7.1 7.08. CLEAR LIQ.HT. POSITION 2, IN. 2.15 2.00 2.05 2.15 2.15 9. 3 1.70 1.80 1.80 1.45 1.50 10. 4 1.80 1.85 1.-85 1.40 1.45 11.:5 1.90 1.90 1.90 2.15 2.15 12. AVG. CLEAR LIQ. HT. IN. 1.89 1.89 1.90 1.79 1.81 13. AVG. LIQ. TEMP. ON PLATE, ~F 100.62 102.2 102.3 99.9 99.9 14. LIQ. TEMP. IN ROT. ~F 99.4 99.6 99.5 92.0 91.8 15. TEMP. OF GAS BELOW PLATE, ~F 106.2 107.9 109.0 104.5 104.5 16. TEMP. IN GAS ROT., ~F 77.5 79.7 80.0 73.0 7..0 17. SAMPLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET 18. TEMP. OF GAS SAMPLE METER, ~F 74.3 74.2 78.0 78.0 78.7 78.5 80.8 80.7 81.8 81.8 19. PRESSURE IN METER IN H20 -- -- -- -- -- -- -- -- -- -- 20. METERED SAMPLE VOL.FT3 1.000 1.000 1.0000 1.0000 1.0000 1.0000 1.000 1.000 1.000 1.000 21. METER CORRECTION FACTOR 0.986 1.003 0.986 1.003 0.986 1.003 0.986 1.003 o.986 1.003 22. CORRECT SAMPLE VOL. FT3 o. 986 1.003 o. 986 1.003 o. 986 1.003 O. 0.986 1.003 O.986 1003 23. LB.MOL.OF GAS/FT3 MET. X 1o5 242.8 242.8 244.5 244.5 241.3 241.3 232.7 232.7 234.5 234.5 24. LB.MOL.OF GAS IN SAMPLE X o105 239.4 244.0 241.3 245.5 238.0 242.0 229.5 234.0 231.5 236.0 25. CYCLOHEXANOL ABSD.IN TUBE, gm 0.3681 0.2934 0.4016 0.3229 0.3960 0.3314 0.3929 0.3022 0.3906 0.2876 26. OYCLOHEXANOL ABSD.IN TUBE, LB. MOL. X 105 0 811 o. 646 0.844 O. 712 0.871 0.730 o. 866 0.666 o. 860 0.633 27. TOT.MOLS.IN SAMPLE LB.MOL.X o105 239.77 244.65 242.1 246.2 238.4 242.73 230.37 234.67 232.36 236.63 28. yo, Y1 MOLS.CYCLOHEXANOL/MOL.GAS 0.003385 0.00oo2648 0.003665 0.00290 0.003662 0.00302 0.00oo378 0.002845 0.003715 0.002685 29. VAPOR PRES. OF LIQ. ON PLATE, IN Hg 0.132 0.133 0.134 0.122 0.122 30. y* 0.00427 -0o. 004308 o. oo004309. 004000 004000 31. EOG% 45.5 54.5 50.0 80.9 78.3 32. NOG = -An (1-EOG) 0.612 0.794 0.696 1.659 1.535 33. vs, FT/SEC 2.17 1.088 1.082 4.44 4.44 34. PG LB/FT3 0.0703 0.0705 O.0704 0.0700 0.0698 35. F = vs PG 0.575 0.289 0.288 1.177 1.173 36. tG, SEC 0. 0944 0.1232 0.1232 0,0997 0. 0972 37. k'Ga, SEC'1 6.49 6.43 5.65 16.65 15.70

TABLE T-G (CONTINUED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA FOR N2-CYCLOHEXANOL SYSTEM) 3.50-IN. WEIR, 4-IN. SPLASH BAFFLE -RUN NO. H-31-A H-31-B H-32-. H-32-B H-33-A H-33-B H-34-A H-34-B 1. GAS ROT. RDG. 80 80 118 118 149 149 46 44 2. LIQ. ROT. RDG. 25 25 25 25 25 25 26 25 3. BAROMETRIC PRESSURE, IN Hg ABS. 29.18 29.18 29.24 29.24 29.24 29.24 29.19 29.19 4. PRESSURE IN GAS ROT., IN Hg 2.30 2.40 2.35 2.20 2.62 2.50 2.40 2.50 5. PRESSURE ABOVE PLATE, IN Hg 1.90 1.90 1.60 1.50 1.60 1.50 2.15 2.20 6. PRESSURE DROP ACROSS PLATE, IN H20 3.80 3.80 4.40 4.40 5.15 5.20 3.30 3.30 7. FROTH HEIGHT, IN. 7.7 7.7 8.8 9.0 100 10 6.7 6.7 8. CLEAR LIQ. HT. POSITION 2, IN. 4.00 4.00 3.90 3.95 3.75 3.80 3.90 3.90 9. 3 3.20 3.20 2.90 2.90 2.70 2.75 3.35 3.35 10. 4 3.40 3.40 3.00 3.00 2.65 2.70 3.50 3.45 11. 5 3.97 3.97 3.55 3.55 3.30 3.55 3.80 3.80 12. AVG. CLEAR LIQ. NT. IN. 3.64 3.64 3.34 3.35 3.10 3.15 3.64 3.63 13. AVG. LIQ. TEMP. ON PLATE,'F 122.0 122.0 122.0 122.0 121.8 121.5 121.7 122.0 14. LIQ. TEMP. ON ROT. ~F 113.0 113.8 118.7 118.9 118.4 118.2 118.2 118.5 15. TEMP. OF GAS BELOW PLATE,'F 139.5 140.5 140.5 140.2 140.0 140 139.0 140.7 16. TEMP. IN GAS ROT., OF 65.1 65.2 71 71.3 70.7 71.1 71.5 ~I.3 17. SAMPLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET 18. TEMP. OF GAS SAMPLE METER, "F 73.0 73.2 72.9 73.0 71.0 71.5 71.3 71.7 70.5 71.5 70.5 71.2 72.0 72-0 72.2 25 19. PRESSURE IN METER IN H0 -- -.- - - - - - -.. 20. METERED SAMPLE VOL. FT3 2.0000 2.-0000 2.0000 2.0000 2.0000 2.0000 2.0000 2.0000 2.0000 2.0000 2.0000 2.0000 2.0000 2.0000 2.0000 21. METER CORRECTION FACTOR 0.986 1.003 0.986 1.003 0.986 1.003 0.986 1.003.986 1.086 1.03 0.986 1.003 0.98686 1.003 22. CORRECT SAMPLE VOL. FT3 1.972 2.006 1.972 2.006 1.972 2.006 1.972 2.006 1.972 2.006 1.972 2.006 1.972 2.006 1.972 2.006 23. LB. MOL. OF GAS/FT3 MET. X 105 242.95 242.95 244.0 244.0 245.3 245.3 245.3 245.3 245.4 245.4 254.4 254.4 244.2 244.2 243.5 243.5 24. LB. MOL. OF GAS IN SAMPLE X 105 479.5 488.0 482.o 490.0 484.0 492.0 484.0 492.0 484.o 492.0 484 492 482 490 480 488 25. CYCLOHEXANOL ABSD. IN TUBE, gm 1.3691 0.5604 1.3594 0.5545 1.4043 0.418 1.4339 0.4092 1.4367 0.4160 1.4513 -- 1.2449 0.5239 1.2392 0.5246 26. CYCLOHEXANOL ABSD. IN TUBE, LB. MOL. X 105 3.01 1.235 2.99 1.22 3.10 0.92 3.16 -0.900 3.16 0.915 3.30 2.74 1.15 2.72 1.153 27. TOT. MOLS. IN SAMPLE LB. MOL.X 105 482.51 489.2 484.99 491.22 487.10 492.92 287.16 492.9 487.16 492.92 487.2 484.74 491.15 482.72 489.15 28. yo, Yl MOLS.CYCLOHEXANOL/MOL. GAS 0.00624 0.002524 0.00617 0.002482 0.00635 0.001867 0.00650 0.001829 0.0065 0.00186 0.00657 0.00186* 0.00565 0.002346 0.00565 0.00236 29. VAPOR PRES. OF LIQ. ON PLATE, IN Hg 0.282 0.282 0.282 0.282 0.28 0.278 0.28 0.282 30. y* o.00908 0.00908 0.00914 o. 00919 0.00907 0.00905 0.00895 o.00900oogooo 31. EOG% 57.5 56.0 61.6 64.3 64.3 65.5 50.0 49.6 32. NOG = -An (l-EoG) 0.859 0.822 0.956 1.03 1.109 1.064 0.0o4 o.69 33. v-, FT/SEC 2.31 2.32 3.42 3.44 4.34 4.34 1.31 1.262 34. pQ LB/FT3 0.0675. 0671 0.067 0.0667 0.067. 0666 0.0682. 0676 35. F vPG 0.60 0.60 0.601 0.885 0.886 1.123 1.12 0.343 0.328 36. tG, SEC 0.146 0.1454 0.133 O. 137 0.1327 0.132 0. 195 0.20e4 37. k'Ga, SEC-1 5.88 5.65 7.19 7.52 8.35 8.08 3.54 3.4 * BORROW FROM RUN H-33-A.

TABLE I-G (CONTINUED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UN'IVERSITT OF MICHIGAN (ORIGINAL DATA FOR N2-CYCLOHEXANOL SYSTEM) 2.-IN. W1EIR, 2.50-IN. SPLASH BAFFLE RUN MO. H-28-A H-28-B -H-27-A H-27-B H1-29-A H1-291. GAS ROT. RDG. 40 40, 80 78 120 120 2. LIQ. ROT. RDG. 25 25 25 25 25 25 5. BAROMETRIC PRESSURE, IN Hfg ABS. 29.58 29.58 29.57 -29.37 29.2 29,5 4. PRESSURE IN GAS ROT., IN H~g 1.90 2.20 2.50 2.50 2.40 2.55 5. PRESSURE ABOVE PLATE, IN Hg1.70 1.90 2.10 2.10 1.64 1.8 6. PRESSURE DROP -ACROSS PLATE, IN 12 0 1.80 1.900 2.45 2.50 5.10 5.2 7.FROTH HEIGHT, IN. 4.53..... 8. CLEAR LIQ. HT. POSITION 2, IN. 2.45 2.50 2.60 2.60 2.55 2.60 9. 5 2.20 2. 20 2.05 2'.10 1.95 1.95 10. 4 2.50 2.50 2.20 2.20 1.90 1.9 1.1. 5 2.55 2.55 2.40 2.40 2.40 12'. AVG. CLEAR LIQ. RT. IN. 2.58 2.34 2.51 2.55 2.20- 2.25 15. AVG. LIQ. TEMP. ON PLATE,'F 121.55 121.55 121.57 -- 121.1 122.61 14. LIQ. TEMP. ON ROT.'F 118.6 119.1 119.0 118.5 119.-7 119.0c%~ 15. TEMP.- OF GAS. BELOW PLATE, OF 154.0 155.2 155.9 155.9 141.2 159 16.'TE P. IN GAS ROT.,'F 80.5 80.4 78.0 78.5 76.2 76.0 17. SAMPLING P0IN OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLE ILE 18. TEMP. OF GAS SAMPLE ETEN,'F 81.6 81.5 81.2 81.5 76.0 76.4 77.5 77.5 76.2 77.2 78.0 8. 19. PRESSUR IN METER IN 1120 - - - -------------------- 20. ETEN CORRECTION FACTOR 2.0205 2.0000 2.0185 2.0000 2.0000 2.0000 2.00o16 2.0000 2.2511 2.0011 2.07 200 21. MTEN CORRECTION FACTOR 0.986 1.005 0.986 1.005 0.986 1.005 0. 986 1.005 0. 986 1.005 0.8.0 22. CORECT SALE VOL. FT5 1.972 2.000 1.9900 2.006 1.972 2.006 1.977 2.006 2.1906 2.0071 1.96.0 25. LB. MOL. OF GAS/FT3 MT. x 1oS 258.75 258.75 258.'7 258.7 242.6 242.1' 241.87 241.85 241.0 241.0 241.1 24. 24. LB., MOL. OF GAS IN SAMPLE x 105 472.0o 479.0 474.0 479 479 486.0 477.6.484 555 484 476.4 48 25. CYCLOHEXAOL ABSD. IN TUBE, gmn 1.2619 0.5700 1'.2457 0.5797 1.5126 0.5502- 1.5402 0.5647 1.6166, 0.4417 1.594 045 26. CYCLOHEXANOL ABSD. IN TUBE, LB. MOL. X ioS 2.780 1.255 2.74 1.276 2.995 1.2,12 294 1.24 55.7.7.9 27. TOT. MOLS.IN SAMPLE LB.MOL. x loS 474.78'480.255 476.74 480.53 482.5 487.2 480.55 485.24 538.56 484.97 479.48 8 28. ~ 1MOLS.CYCLOREXAOL/MOL. GAS 0.00585 0.00262 0.00575 0.00266 0.00600 0.002492 0.00o614 0.00256 0.00)663 0.002000 0.004C.00 29, VAPOR PRE-S. OF LIQ. ON PLATE, IN Hg 0.278 0.278 0.274 0.274 0.,29 0.28 3o. Y* 0.00894 0.00895 0.00870 0.00870 0.00959 0.005 51. EOG% 51.1% 49.5% 56.5 58.5 62.7 60.0 52.' NoG0 -An (1-EOG) 0.715 0.68 0.834 0.90 0.985 0.91 55. VS, FT/ SEC 1.110 1.136 2.26 2.19 5.42 3.44 34. PG LB/FT3 0.0676 0.0678 0.0687 0.0o69 o.o68 0.067 55. F = VS4TG 0.2886 0.296 0.5925 0.575 0.891 0.89 56. t1~ SEC 0.148 0.1365 0.1205 0.128 0.1065 0.10 57..kqGa, SEC-1 4.82 4.98 6.-67 7.000 9.55 8.91

TABLE I-G (CONTINUED) PLATE EFFICIENCIES IN RECTANGUmLAR COLUMN AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA FOR N2-CYCLOHEXANOL SYSTEM) 2-IN. -WEIR, 2.50-IN. SPLASH BAFFLE RUN NO. H-30-A H-O-B H-14-A H-14-B H-25-A H-25-B 1. GAS ROT. RDG. 150 150 40 40 40 40 2. LIQ. ROT.RDG. 25 25 25 25 25 25 3,. BAROMETRIC PRESSURE, IN Hg ABS. 29.10 29.10 29.26 29.26 29.51 29.51 4. PRESSURE IN GAS ROT.,IN Ng 2.65 2.70 1.60 1.70 2.40 2.20 5. PRESSURE ABOVE PLATE, IN Hg 1.64 1.74 1.50 1.55 2.15 2.-05 6. PRESSU DROP ACROSS PLATE, IN H2O 4.18 4.05 2.00 2.00 2.00 1.95 7. FROTH HEIGHA, IN. 7,5 7.5 4.6 4.7 4.6 4.60 8. CLEAR LIQ. H. POSITION 2,IN. 2.50 2.60 2.50 2.50 2.50 2.50 9. 3- 1.95 1.95 2.20 2.20 2.15 2.20,0. 4 1.90 1.95 2.27 2.30 2.25 2.25 11. 5 2.55 2.60 2.33 2.35 2.30 2.350 -12. AVG. CLEAR LIQ. W. IN. 2.25 2.28 2.33 2.34 2.350 2.31 13. AVG. LIQ. TEMP. ON PLATE, F 121.82 122..36 99.32 99.1 101.0 100.9 14. LIQ. TEMP. ON ROT.'F 119.0 119.5 95.9 95.2 99.7 99.0 I 15. TEMP. OF GAS BELOW PLATE,'F 138.1 138.2 102.6 10o2.6 106.7 114.2 16. TEMP. IN GAS ROT., OF 75.1 75.7 75.8 75.6 75.0 75.5 17. SAMPLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INT 18. TEMP. OF GAS SAMPLE METER, ~F 74.1 75.5 75.1 76.9 76.8 76.2 76.8 76.2 71.0 70.9 71.5 72.0 19. PRESSUREIN METER IN Mo.0.......... 20. METERED SAMPLE VOL. FT3 2.06 2.0000 2.0000 2.0000 1.0000 1.0000 1.0000 1.0000 2.0000 2.0000 2.0000 21. METER CORRECTION FACTOR 0.986 1.003 0.986 1.003 0.986 1.003 0.986 1.003 1.003 0.986 0.986 1.003 22. CORRECT SAMPLE VOL. FT3 1.980 2.006 1.972 2.006 o.986 1.003 o.986 1.003 2.006 1.972 1.972 2.06 23. LB. MOL. OF GAS/FT3 MET. X 105 242.7 240.7 241.3 232.05 242.1 242.3 242.1 242.3 248.0 248.0o 247.0 247. 24. LB. MOL. OF GAS IN SAMPLE X 105 481 484 476 466 287.3 243.0 238.7 243.0 497.0 489.5 487.7 495 25. CYCLOHEXANOL ABSD. IN TUBE, gm 1.3998 0.4688 1.3999 0.4454 0.,3385 0.2535 0.3426 0.2424 0.7650 0.4942 0.7494 0.496 26. CYCLOHEXANOL ABSD. IN TUBE, LB. MOL. X o105 3.08 1.0531 3.08 0.98 0.746 0.558 0.755 0.535 1.683 1.089.65 1.092 27. TOT. MOLS. IN SAMPLE LB. MOL. X 105 484.08 485.-031 479.1 466.98 239.45 243.6 239.5 243.5 298.7 490.6 489.55 496.1 28. Yo, Y1 MOLS. CYCLOEXANOL/MOL. GAS 0.00636 0.002134 644.1 0.00 03115 0.002291 0.003153 000197 0.0038 0.0022 0.003375 0.0001 29. VAPOR PRES. OF LIQ. ON PLATE, IN Hg 0.2-8 0.282 0.127 O0.127 0.137 -0.135 30. y 0.00911 0.00915 0. 00413 0.00142 0.00435 0.00,43 33. EoG% 60.6 61.5 44.93 49.71 54.5 59.6 32. N00 = -n (1-E00) 0.933 0.955 05 13 0.730 0.789 O.91 33. vt8, FT/SEC 4.36 4.35 1.107 1.082 1.08 1.084 34 PG LB/T 0.0667 0.067 0.0705 0.0708 0.0718 0.0718 35. F = Vs4 PG 1.27 1.128 0.294 0.288 0.2898 0.2904 36. tG, SEC 0.103 0.10 0.1711 0.1815 0.1775 0.176 37. k' a, SEC-1 9.30 9.55 3.58 4.022 4.44 5.17

TABLE T -G (CONTINUD) PLATE EFFICIENCIES IN RECTANGULAR COLUMNt AT UNIVERSITM'OF. MICHIGAN (ORIGINAL DATA FOR N2-CYCLOBEXANOL SYSTEM)..,- 2-IT. WEIR, 2.50-IN. SPLASH BAFFLE Ru~ Noo ~~~ ~ ~ ~~~~~~~~ ~ ~~~~~~H-12-A H-12-B.3-A -1-B H -A- H-r1T-A 1. GAS ROT. RDG.; 80 80 80 80 80 80 2, LiQ'. ROT. RDG. 25 25 ~25 -25 25 25 3, BAPOMT-RiC PRESSUYRE, IN HEg: —ABS. 2 8'~84 28.84 29,27 29027 29,'37 29.66 4. P IESREE IN GAS ROT. iN Hg 2.1 2.20 2.352-3 2.25...25 5-~'PHE~Sv.' AB0:V-E PLATE, IN Hg 1. 0. 9 1 9 5 1$i8522 6o,- PKESSUP~E. DROP ACROSS PLATE, iN B20 2.60 2,40 2.65 2.40; 2.40 2,6 7o FROTH BEIGHT, IN. 4.3 4,5 5.3 5.2 5o505! 8. CLEAR. ILIQ. HT. POSITION 2, IN. 2. 5. 6 2 6 5 22.2.6-0 2 c),.. 3 ~ ~ ~ ~ ~ ~~~ ~ ~ ~ ~~~~ ~~~~2.10 2.0 2.;i0 2.07.' 2.10.0 "O....'' 4 2,~~~~~~~~~~~~~~~~220 2,20 2,5.2!20 ~' 2,i ii, ~ ~ ~~~~~~~~ ~ ~ ~~5. 2,. 4.0 2.40 2,,40 2,, 92,40 2, 12, AVG. CLE.3YR LIQ. HT.,IN. 2.34- 2 31 2. 3i5 2.321 2 33 2 2 35 CG I. T E. ONPA E F 99.9'7 99.5' 99 ~ 99,.2!0, i%~ ~.'-g.'TEYW. ON ROT. ~F 97.2 97.0C:91.0 90c9.9' 95.2h 5.TE~, M0 F 67 AS:BELOW PLATE, ~F t10l4.8.10. 103. O2,Z10, 98,6' l1'.' -TEMP, iI{ GAS ROT., ~F 73- 0.73.2,68,2 68,,0 6609 69.:17. SADTLi-NG POINT OUTL ET INLET OUTLET INLET ~OULET! ET 0Uf INE U TiLET' b'-'E i:U~ 13. TMEMP. OF GAS SAMPLE METER, OF 69.0 69.0 69.0 69.0 75~2 74.0 75,8, 474' 70.6 70o -7 59.0'94 19,PFRESSOP17 IN METER IN Hp 0..............'20,, }vf,~,/ ED S.]W~IPTF VOL. FT3 i,.0000!. 0004 1. 0000 ]L, 0000) 1!C 100 1.0-0 10 "2i, MsTER'C0'1CTiON FACT0OR. 986 103 Oyo.6 O 0,',8 03 0.06 " 0,-~..,9~'22-,..'COPRRBCT SAMPLE, VOL. FT.-:0oo 98-03 O. 986 I. 00- 0.56 l- 003 O.9:,:, l 05 0,9:.('50,96:.(;?5!3. B MOL. OF GAS/PT3 ME;X 105 243.8 243.8 243.8 243.8 2kh7 } 24 8,5 246.9283 24.~ IS. MOLo- OF GAS IN SAMPLE X 105 240.5 244.5 240.5 244.5'243~8 2f49,2:245~,.8 2149,6 243'~2 2;~o.247.9 25'25~. CYCLOHE-XANOL ABSD. IN TUBE, g 0.3077 0.2599 0.3064 0.2225.53 0.2528 0.3150 0,,2570 0.3813 0.2469 ~0.5ri2 2:; 26. CYCLOHEXANTOL ABSD. IN'TUBE, LB. MOL.- X 10o5 0.678 0.572 0.676 0.49i1 0-7 56 0.556 0,69 0L5 66( 0,814, 0,5 4`4 o,6850E~5 27. TOT. M OLS. IN SANPL LB. MOL. X 105 241.2 245 07.241.2 246. 0 24-4.6 24 9.1 i 244 5 250.2.244 0 248.4 248W6 o 2.6.o,, Yl... 28~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. v2v.2' 7 po 29. VAPOR PRES. OF LIQ. ON PLATE, INT _wg O. 128 0. 127 0o7. k~7 0,!0.0, 1 Y*0 "'0.0,0417 O., 00413 0.004.6 6,.0:08'.01,004 16'.. 31. EOG%26239 GG~ ~~~~~~~2. 5. 6. 2,9, 63.6 4. 32. NOG = -in (1-EoG) 0.300 0.476 0.659 0.35 1.411" 0.746 33. vs, FT/SEC 2.15 2.11 2.177 2,18 2.18 21 34. PG O.O-fo~ ~ ~~.074 0.07o4, 0.0713 0.0715 0.0715oo3'5 5-~ F =v?' PrG O. 570 O. 558:.,0.581. O. 583 0,583 O. 58 36. tG oE.0665 o 0 o766 o'1532 O.ilo11 ~ O 3.kGaI SEC-! 37. -kt~~a, SEC-' ~4.51 6.21. 5.82 3,18 8.34 6.85

TABLE Z-G (coNThUmp) ~r~ E~TC~CES IN ECT~mmAR cornram AT UN~SnW OF ~C~GAN (0RTGINAL DATA FOR N2-~CnO~X~OL SZS~) 2-IN. WEIR, 2.50-IN. SPLASH BAFFLE RUN NO. H- L7-B H-21-A H-21-B E-15-B H-11-A H-11-B E-18-A 1. GAS ROT. RDG. 80 80 80 120 120 120 120 2. LIQ. ROT. RDG. 25 25 25 25 25 25 25 3. BAROMETRIC PRESSURE, IN Hg ABS. 29.66 29.4 29.4 29.$7 29.22 29.22 29.56 4. PRESSURE IN GAS ROT., IN Hg 2.20 2.55 2.40 2.70 2.50 2.60 3.05 5. PRESSURE ABOVE PLATE, IN Hg 1.80 2.17 2.10 3.20 1.90 1.90 2.50 6. PRESSURE DROP ACROSS PLATE, IN H20 2.50 2.43 2.40 2.00 3.10 3.10 3.10 7. FROTH HEIGP~, IN. 5.1 5.3 5~4 6.2 6.3 6.4 6.1 8. CLEAR LIQ. HT. POSITION 2, IN. 2.60 2.60 2.60 2.40 2.50 2.50 2.50 9. 3 2.05 2.05 2.10 1.80 1.85 1.85 1.80 10. 4 2.20 2.20 2.20 1.80 1.85 1.85 1.80 ~. 5 2.35 2.40 2.40 2.50 2.35 2.30 2.30 12. AVG. CLEAR LIQ. HT. IN. 2.30 2.31 2.52 2.08 2.14 2.12 2.10 13. AVG. LIQ. TEMP. ON PLATE, ~F 95.8 98.9 99.9 100.3 99.7 99.7 96.2 14. A LIQ. TEMP. ON ROT. ~F 94.9 96.9 98.1 90.8 97.8 97'6 96.1 (DO 15..TEMP. OF GAS BELOW PLATE, OF 101.2 103.1 103.3 105.8 10~.5 104.9 103.1 _~ 16. TEMP. IN GAS ROT., ~F 71.8 72.2 73.3 66.0 73-5 74.3 70.4! 17. SAMPLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTIET INLET OUTLET INIET OUTLET INLET 18. TEMP. OF GAS SAMPLE METER,'F 69.0 69.O 71.5 67.7 68.8 71.8 71.7 71.5 67.9 67.9 67.9 674 9 68.4 68.3 19.PRESSURE IN METER IN T~0........................ __ -. 20. METERED SAMPLE VOL. FT 1.0000 1.0000 1.0000 1.0025 1.0000 0.9975 1.OlO 1.000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 21. METER CORRECTION FACTOR 0.986 1.003 1.003 0.986 0.986 1.003 0.986 1.003 0.986 1.003 0.986 1.003 0.986 1.003 22. CORRECT SAMPLE VOL. FT3 O. 986 1.003 1.003 0.990 O. 986 1.000 O. 996 1.003 O. 986 1.003 0.986 1.003 O. 986 1.003 23. LB. MOL. OF GAS/FT3 MET. X lO5 251.1 251.1 256.3 249.5 248.4 246.2 245.5 245.5 248.0 248~0 248.0 248.0 250 250 24. LB. MOL. OF GAS IN SAMPLE X 105 247.9 252 247 246.6 245.2 246.2 244.2 246 244.2 248.6 244.2 248.6 246.2 250.0 25. CYCLOHEXANOL ABSD. IN TUBE, gm 0.3312 0.2198 0.3530 0.2122 0.3562 0.2051 0.3858 0.2376 0.3506 0.2755 0.3359 0.2751 0.5298 0.2098 26. CYCLOHEXANOL ABSD. IN TUBE, LB. MOL. X 105 0.729 0.484 0.776 0.468 0.785 0.451 0.849 0.522 0.773 0.606 0.735 0.605 0.725 0.461 27. TOT. MOLS. IN SAMPLE LB. MOL. X 105 248.6 252.5 247.8 247.1 246.0 246.7 245.05 246.5 244.97 249.2 244.9 249,2 246.9 250.5 28. Yo, Yl MOLS. CYCLOHEXANOL/MOL. GAS 0.00293 0.00191 0.00314 0.001891 0.00319 0.001828 0.00346 0.00212 0.00315 0.002436 0.00300 0.002428 0.00294 0.00184 29. VAPOR PRES. OF LIQ. ON PLATE, IN Hg 0.111 0.126 O.128 0.130 0.128 0.127 0.120 30. y* 0.003526 0.003990 O. 00401 O. 00415 O. 00412 O. 00408 O. 00376 31. EOG% 63.1 59.5 62.5 66.0 42.3 34.5 57.3 32. N0G =-~n (1-EoG) 0.996 0.905 O.981 1.079 0.548 0.424 0.851 33 vs, FT/SEC 2.146 2.13 2.15 3.28 3.28 3.30 3.24 34. PG LB/FT3 0.0725 0.0725 O.0721 0.O719 O.0713 O.0710 0.073 35. F = vjp~ 0.576 O. 574 O. 577 0.879 0.876 O. 880 0.875 36. tG, SEC'~ 0.1087 O. ll7 O. l191 O. 1050 O. 106 O. 108 O. 1029 37. k'Ga, SEC-1 9~19 7.74 8.25 10.29 5.19 3.93 8.28

TABLE I-G (CON2TIIUED) PLATE EFFICTENCIES IN SECTANGULAR CGLUUME AT r10J4RSTTV. OF' MICHEGAB (ORIGINAL-DATA FOR N2-CYCLOHEXANOL SYSTEM) 2-IN. SETS, 2.50-IN.SPLASHRBA LE RUN NO. H1-19-A H1-19-S 11-20-A 11-20-S H-22 -A 11-22 -B 11-101. GAS ROT. RDG. 120 120 12 10O2 120 1.60 2. LIQ. ROT. RDG. 25 25 25 25 25 25 2 3. BAROMETRIC PRESSURE, IN Hg ABS. 29.55 29.53 29.4 29.4 29.2 29.2 29 1 4. PRESSUR IN GAS ROT., IN Hg 2.70 2,50 2.40 2.50 2.40 2.40 530 5. PRSSUR ABOVE PLATE, IN Hg9 2.10 1.90 1.70 1.90 1.70 1.70 1 6. PRSSUR DROP ACROSS PLATE, IN 1120 5.00 2.10 2.20 2,50 2.90 2.90 4 20 7. FROTH IEIGHT, IN. 6.6 6,5 6.4 6,5 6.5 6.5 8. CLEAR L14. 111. POSITION 2, IN. 2,50 2.50 2.50O 2,50 2.50 2.50 2 9. 3 1.90 1.90 1.80 1,80C 1.80 1. 80 1 8 10. 41.90 1.90 1.80 1.80 1.80 1.80 8 II. 5 2.55 2.50.2.30 2,50 2,510 2,50 20 12. AV-G. CLEAR LIQ, NT. IN. 2.16 2.15 2.10 2.10 2.10 2.10 2 15. AVG. LI14. TEMP. ON PLATE, -F 99.0 909.9 99.5 99.9 99,6 99.7 ia 14 LIQ. IBM1-P ON ROT.,'F 91.0 (2.4 97 1079 595 955 1). TEMP. OF (:-A BELOW PLA-TEE,'F 104.5 1o4. 9 104.8.05 1o 106 5 i6, TEMP. I- GAS ROT.,'F 72 71 9 71.7 81. 1 L 71.8 67,1 I" S1AMPLING -PUIN OULTLE T INLET OUTIET INLET 00TLET 11N\.IRT 001271 I NTR.o OUTIET ONTTLET~ INTE 18~ IEMP. OF GA SAMPLE METER, -F 84. 6 84,1 85. 4 85 0 75.0 72 4''1 765 6. 76. 764 10. lRESURF 1 SETER RE 1120 — - 20O =ERED 0 AIIILE VOL. P15 1.100 1.100 1. 0OC0 1 (0000 1 CCO 000 01%0 1(9-0 i~~l" 170 2 7 TZ0" 012 1271CRIO FACTOR 0,986 1,005 0. 086 1.005 1.005 o ~ 86 o,86 0.986 07 O.C ) 1 22~ COJRR~0T SAP12fLE V01. FT5 1.0846 1.:L03 C 9806 1.00 1.~005 0 5 96 I cR 6 0)78 TO i617 RE3 MOL /T0 ST7 1- 4. 246. 5 256. L 245,2 24 ~ 2h 211 1, 24- T3 MOL 8F GAS IN SAMPL x 105 267.1, 271. 250 27 2458 2 4 4oR o 4a L11404 2-_.'CLOEEXARDTL ARGO. IN TUBE, g 0.5;972 0,2415 0 T503 0 1 o41 0.57751 j 0,1255 08 53-,e' 210 ~- 26 CYLOE"'72OL AR20. INTUBE, T-B. TOL, X 8y 0,51 075 04585 0. 47 1 - 0 5; >50 0.-478 27. TOT ROJINS1E REB. 1100. X 15 267(,97 272.2 252. 257.9c 24688 24. 2 ~6. 4(7.42 412.8 4-u,;.7 8 28. ml L LOREXAROL/1EOL. GASB ) 0o,005.26 0.001-,944 [BO51 0.-00180 0.005542 0.010112074 0,0558 0.0182 C,,536c 3 29. >202'O PROS. OF LIQ. OR 120 N 11g 0,12-8 0.129 0. 27 0 1"1-2 0.12",70 2.10' 0,00o~414 0.004-16 0.00o40c8 0.C 004110 O c oo'~~~~~~~~~~~~~~~ ~~~55,-5 68.5 66.2 6756 52 N --'2 0,809 1154 1,.085 0 2, 1.261 7 PT58~C3.52 5.52 5.28 A26 3.29 (71 00B1/FT3 0.07075 0.0'71 0.0714 0 0~i60.7v0 o 55. F =vP,0.881 0.88~5 0.876 0,~872 0.875 0. 872 1 56. tSE C 0,150.109 0.109 0. 11220.15.18009 G7~a, EC 10.0 10.6 9.-93.11..08

TABLE I -G (CONTNED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY -OF MICHIGAN (ORIGINAL DATA FOR N2-CYCLOHEAOL SYSTEM) 2-IN. WEIR, 2.50-IN. SPLASH BAFFLE RUN NO. H-10-B H-23-A H-23-B H-24-A H-37-A H-37-B H-36-A 1. GAS ROT. RDG. 160 160 160 160 51 51 -96 2. LIQ. ROT. RDG. 25 25 25 25 50 50 50 3. BAROMETRIC PRESSURE, IN Ng ABS. 29.1 29.21 29.21 29.53 29.50 29.50 29.38 4. PRESSURE IN GAS ROT.,IN Bg 2.80 3.20 3.10 3.00 2.-90 2.85 2.80 5. PRESSURE ABOVE PLATE, IN 9 1.70 2.10 2,05 1.95 2.60 2.57 2.30 6. PRESSURE DROP ACROSS PLATE, IN H20 4,20 4.45 4.45 4.20 2.90 2.85 3.60 7. FROTH HEIGEr, IN. 7.5 7.8 7.8 7.8 6.2 6.4 7.8 8. CLEAR LIQ. HT. POSITION 2, IN. 2.50 2.40 2.50 2.50 3.40 3.40 3.55 9. 3 1.8o 1.75 1.8o 1.85 2.8o 2.80 2.65 10. 4 1.80 1.70 1.80 1.80 2.95 2.95 2.70 11. 5 2.50 2.40 2.45 2.50 3.20 3.20 3.30 12. AVG. CLEAR LIQ. HT. IN. 2.15 2.o6 2.14 2.16 3.09 3.09 3.05 1 13. AVG. LIQ. TEMP. ON PLATE, ~F 100.5 99.9 99.2 99.5 100.5 100.2 100.2 14. LIQ. TEMP. ON ROT.,~F 90.8 95.3 95.1 97.8 99.8 99.5 99.6 OO 15. TEMP. OF GAS BELOW PLATE, ~F 113.0 105.5 105.0 105.5 109.58 110.1 110.3 k, 16. TEMP. IN GAS ROT., ~F 68.1 72.7 72.5 71.2 76.06 76.1 72.5 1 17.~ SAMPLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUT LET 18. TEMP. OF GAS SAMPLE METER, ~F 73.8 73.8 76.0 75.8 77.0 77.5 64.5 64.o 98.1 77.0 98.3 77.0 75.3 75.0 19. PRESSURE IN METER IN 20 20. METERED SAMPLE VOL. TT3 1.0000 1.0000 2.000 0 2.0000 2.0000 2.0007 2.0000 2.0000 2.0000 2.0018 2.0040 2.0014 21. METER CORRECTION FACTOR o.986 1.003 o.986 1.003 1.-003 o.986 1.003 o.986 O.986 1.003 O.986 1.003 o.986 1.003 22. CORRECT SAMPLE VOL. FT3 o.986 1.003 1.972 2.oo6 2.006 1.972 2.016 1.972 1.'972 2.oo6 1.974 2.0100 1.973 23. LB. MOL. OF GAS/FT3 ME7. X 105 242.75 -242.75 241.7 241.8 241.2 240.8 252.8 252.9 242.86 243.65 242.7 243.7 243.8 244.0 24. LB. MOL. OF GAS IN SAMPLE X 105 -240 244 476.2 484.6 484 476 510.0 499.o 478.92 488.76 479.1 489.7 481.1 25. CYCLOHEXANOL ABSD. IN TUBE,gm 0.3656 0.2867 0.7887 o.4698 0.8194 0.4651 o.8404 0.4567 0.6756 0.3393 o.6673 0.3352 0.7938 26. CYCLOHEXANOL ABSD. IN TUBE, LB. MOL. X 105 0.804 0.631 1.734 1.003 1.80 1.022 1.85 1.0o4 1.487 0.747 1.4692 0.738 1.7477 27. TOT. MOL. IN SAMPLE LB. MOL. X 105 240.8 244.6 477.9 485.6 485.8 477 511.85 500.0 480.4 489.5 480.53 490.5 482.85 28. yo, Yl MOLS. CYCLOHEXANOL/MOL. GAS 0.00334 0.00258 0.00363 0.002128 0.00371 0.002142 0.00362 0.00201 0.003095 0.001526 0.003057 0.001505 0.0036 29. VAPOR PRES. OF LIQ. ON PLATE, IN H9 O. 132 O.128 0.127 O.128 0.131 0.130 0.131 30. y* 0.00429 O.004118 O.00406 o.oo4o O.o004081 O.004054 31. EOG% 44.4 75.5 81.7 80.2 61.41 60.89 80. 44 32. NoG =-_n (1-EOG) 0.588 1.406 1.70 1.62 0.951 0.937 1.631 33. VS, FT/SEC 4 * 77 4.37 4.24 4.30 2.73 2.75 3.641 34. PG L0/F93 o.o6g8 O. 0715 0.072 O.O73 O0 0755 -0-07539 0.074 35. F = Vsq PG 1.261 1.167 1.136 1.161 o. 754 O0 755 O 994 36. tG, SEC O. 0935 0.1093 O. lll o. 114 o. 0950 0.1005 O 1087 37. k'Ga, SEC-1 6.29 12.85 15.31 14.2 10.01 9.32 15..1

TABLE I-G CONTINUED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN (ORIGINAL DATA -FOR N2-CYCLOREXANOL SYSTEM) 2-IN. MSIRi, 2.50-TN. SPLASH BA-7LE RUN NO. H-36-B H-55-A H-35-B H-38-A H-38-B H-39-A H-39-B 1. GAS ROT. RDG. 96 100 100 100 100 147 1)47 2. LIQ. ROT. RDG. 50 50 50 50 50 50 50 3. BAROMETRIC PRESSURE, IN Ag ABS. 29.38 29.29 29.29 29.50 29.50 29.30 29.30 4+. PRESSURE IN GAS ROT.,IN Hg 2.80 3.00 5.00 2.80 -2.85 3.30 3.30 5. PRESSURE ABOVE PLATE, IN Hg 2.25 2.50 -2.40 2.530 2.30 2.20 2.22 6. PRESSURE DROP ACROSS PLATE, IN 1120 3.6o 3.80 3.70 5.60 2.70 4.-60 4.65 7. FROTH REIGHI, IN. 7.8 7.7 7.7 7.8 7.6 9.5 9.5 8. CLEAR LIQ. HT. POSITION 2, IN. 3.55 3.55 3.55 3.55 5.55 5.40 3-.40 9. 3 2.65 2.62 3562 2.65 2..60 2.45 2.45 10. 4 2.70 2.65 3.65 2.65 2.65 2.50 2.50 11. 5 5.30 3.25 5.30 3.25 5.25 5.20 3.20 12. AVG. CLEAR LIQ. HT. IN. 3.05 5. 02 3.03 3.03 3.05 2.84 2.84 13. AVG. LIQ. TEMP. ON PLATE,'F 100.4 100.3 100.4 100.4 100.4 00o.4 100.4 14. LIQ. TEMP. ON ROT. OF 99.7 97.9 98.2 99.7 99.5 98.o 97.9 15. TEMP. OF GAS BELOW PLATE,'F 109.85 106.5 104.0 113.5 115.7 109.1 108.5 o L6. TEMP. IN GAS ROT.,'F 72.7 72.4 72.4 71.6 71.3 75. 6 75.6 1.7. SAMPLING POINT OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET INLET OUTLET IET 18. TEMP. OF GAS SAMPLE METER, OF 75.6 75.5 78.0 77.9 78.0 78.8 77.5 76.0 77.1 76.0 80.5 80.5 80.5 80.5 19. PRESSURE IN METER, IN 2O20. METERED SAMPLE VOL. FT3 1.9989 2.0000 2.0000 2.0085 2.0000 1.9917 2.0000 2.0000 2.01-40 2.2002 2.0000 2.0000 2.0010 2.0000 21. METER CORRECTION FACTOR 0.986 1.003 0.-986 1.003 0.986 1.005 0.-986 1.005 0.986 1.003 o.986 1.003 0.986 1.005 22. CORRECT SAMPLE VOL FT3 1.971 2.006 1.972 2.015 1.972 2.0000 1.972 2.006 1.9858 -2.28-36 1.972 2.006 1.973 2.006 23., LB. MOL. OF GAS/FT3 MET. X 105 243.6 243.8 244.5 244.5 241 241.0 243.3 244.4 243.58 244.36 239.2 239.2 239.2 239.2 24. LB. MOL. OF GAS IN -SAMPLE X 105 48o.o 489.0 482 495 476 482.0 479.8 490.2 485.70 539.70 471.7 479.8 472.0 479.8 25. CYCLOREXANOL ABSD. IN-TUBE, gm 0.7059 0,3330 0.7577 0.3123 0.6467 0.3038 0.6479 0.3056 0.64-57 0.3309 0.6898 0.5272 0.6881 0.5296 26. CYCLOREXANOL ABSD. IN TUBE, LB. MOL. x lo5 1.554 0.7332 1.67 0.688 1.423 o.6685 1.4265 0.6728 1.422 0.7285 1.5177 0.7199 1.5139 0.725k 27. TOT. MOLS. IN SAMPLE LB. MOL. X lo5 481.55 489.73 482.76 495.7 4.77.4 482.7 - 481.23 490.87 485.12 540.43 473.22 480.52 473.5 480.55 28. Y Y1 0.003227 0.001500 0,003470 0.001395 0.002978 0.001385 0.002960 -0.001371 0.002951 0.001348 0.005207 0.001498 0.003150 29. VAPOR PRES. OF LIQ. ON PLATE, IN Hg 0.132 0.130 0.130 0.132 0.132 0.130 0.130 30. y* 0.004173 0.00408 0.00410 0.004151 0.004151 0.004162 0.004184 31. EOG% 64.65 77.3 58.7 57.16 56.47 64.15 65.10 - 52. NOG = -In (1-E0G) 1.040 1.484 0.885 0.8475 0.8303 1.o26 0.9955 33. vs, FT_/SEC.3.646 3.662 3.675 3.717 3.72 4.616 4.62 34. PG LB/FP3 0.0744 0.0751 0.0748 0.07402 0.0741 0.0741 0. 0740 35. F = vp0 0.9917 1.007 1.003 1.007 1.007 1.251 1.256 36. tG, SEC o-5786 0.1066 0.1059 0-.1071 0.1026.0.1205 0.1203 37. k-'0a, SEC- 9.576 13.95 8.357 7.913 8.095 8.53 8.28

TABLE II-G PREDICTION OF VAPORIZATION DATA NG Percent NG Percent NG Percent NG Percent Weir Predicted Deviation Predicted Deviation Predicted Deviation Predicted Deviation Height Liquid NG by Equation Equation by Equations Equations by Equation Equation by Equation Equation System Run No. Inches Rate, gpm Experimental (123) (113) (14)+(124) (114)+(124) (109) (109) (108) 5 108) N2-C2H4Br2 E-7A 1-1/2 7.9 0.489 0.565 13.53 -0.533 8.25 O0.794 62.40 0.627 28.35 E-7B 0.588 0.560 - 5.06 0.529 -11.21 0.785 33.55 O.618 5.06 E-SA 0.696 0.586 -1884 0.557 -24.97 0.793 14.98 0.672 3.43 E-8B 0.677 -0.610 -10.95. 580 -16.71 0.818 20.90 0.694 2.55 E-9A 0.678 0.596 -13.79 0.566 -19.80 0.810 19.40 o.667 1.57 E-9B -0.752 0. 594 -26.63 0.564 -33.33 0.808 7.42 0.665 11.51 E-1OA 0.692 0.578 -19.80 0.548 -26.27 0.800 15.59 0.635 8.20 E-1OB 0.672. 595 -13.00 0.564 -19.16 0.818 21.75 0,650 3.34 E-11A 0.734 0.676 - 8.61 0.644 -14.06 0.910 23.95 0.685 6.'72 E-llB 0.754 0.679 -11.07 0.645 -16.86 0.915 21.42 0.691 8.30 E-12A 3-1/2 7.9 0.821 1.036 20.76 0.986 16.7 1 53.60 0.944 15.00 E-12B.0.917 1.023 10.33 0.973 5.75 1.250 36.30 0.934 1. 85 O E-13A 1.037 0.924 -12.29 0.877 -18.25 1.144 10.29 0.904 12.84 E-13B 1.051 O. 926 -11. 36 O. 879 -17.30 1.146 11.12 O. 906 12.10 E-14A 1.012 0.939 - 7.82 0.893 -13.32 1.143 12.97 0.933 7.79 E-14B 1.028 0.930 -10.52 0.885 -16.19 1.136 10.49 0.928 9.69 E-15A 0.914 1.127 18.92 1.070 14.55 1.355 48.25 1.007 10.16 E-15B 1.070 1.160 7.76 1.101 2.84 1.384 29.30 1.027 4.02 E-16A 0.920 0.946 2.79 o.899 - 2.32 1.147 24.65 0.956 3.94 E-16B 0.906 0.928 2.36 0.881 - 2.88 1.129 24.65 0.952 5.04 E-17A 1.014 0.924 - 9.78 0.876 -15.77 1.153 13.75 0.893 11.90 E-17B 0.920 0.947 2.85 0.899 - 2.35 1.175 27.67 0.910 1.08 E-18A 15.9 0.899 1.405 36.00 1.333 32.55 1.598 77.80 1.185.31.78 E-18B 0.888 1.324 32.94 1.255 29.26 1.530 72.40 1.140 28.35 E-19A 0.978'1.084 9.75 1.028 4.91 1.292 32.01 1.022 -4.51 E-19B 0.932 1.089 14.46 1.036 9.99 1.295 38.90 1.024 9.86 E-20A. 933 0.953 2.07. 902 - 3.46 1.165 24.80 0.956 2.42 E-20B 0.927 0.972 4.59 0.920 - 0.73 1.181 27.40 0.969 4.52

TABLE II-G (CONT'D) PREDICTION OF VAPORIZATION DATA NG Percent NG Percent NG Percent NG Percent Wei-r Predicted Deviation Predicted Deviation Predicted Deviation Predicted Deviation Height Liquid DIG by Equation Equation by Equations Equations by Equation Equation by Equation Equation System Run No. Inches Rate, gpm Experimental (123) ~123) (114)+(124) (114)+(124) (109) (109) (108) (108) N2-Cyc lohexanol H-3A 1-1/2 8.0 0.1)21 1.020 9.70 O.966 4.70 O.948 2.90 0G. 352 1.20 H-3B 1.061 0.780 -36.03 0.736 -44.15 0.777 26.80 0.769 27.55 H-7A 0.995 0.727 -36.88 0.690 -44.29 0.786 21.00 0.753 24.38 H-7B 0.612 0.714 14.35 0.679 9.92 0.776 26.70 0.742 21.19 H-8A 0.794 0.594 -33.78 0.563 -41.01 0.877 10.50 0.789 0.63 H-8B 0.696 0.592 -17.49. 562 -23.83 o.878 26.10 0.790 13.49 H-28A 2 8.0 0.715 0.819 12.75 0.782 8.60 0.972 35.95 Q. 912 27.60 11-28B 0.680 0.789 13.80 0.750 9.40 0.943 38.65 0.8 93 31.355 H-27A o. 884 O. 997 11. 33. 947 6.66 O. 911 3.02 0.917 3.70 H-27B 0.900 1.000 10.06. 949 5.23 0.923 2.54 0. 924 2.66 H-29A O. 985 1.049 6.14 1.003 1.82 0.829 15.88. 862 12.44 H-29B 0.917 1.045 12.28 O. 998 8.16 o.830 9.53 0.871 4.98 H-30A O. 933 1.135 17.80 1.087 14.17 0.824 11.70 O. 881 5.55 H-30B 0.955 1.135 15.80 1. 085 12.00 0.823 13.82 o. 882 7.66 H-25A 0.789 0.764 - 3.29 0.720 - 9.53 1.076 36.40 O. 963 22.10 H-25B 0.910 0.761 -19.56 0.718 -26.74 1.072 17.75 0.959 5.35 H-21A 0.905 0.823 - 9.94 0.776 -16.64 0.876 3.16 0.835 7.77 H-22A 1.126 1.016 -10.85 o.965 -16.72 0.858 23.80 0.846 24.81 H-22B 1.082 1.017 - 6.42 0.966 -12.02 o.860 20.50 0.848 21.60 H-31A 3-1/2 8.0 0.859 1.130 23.96 1.078 20.31 0.994 15.90 1.002 16.63 H-31B O. 822 1.141 27.95 1.083 24.13 1.006 22.40 1.017 23.80 H-32A 0.956 1.252 23.67 1.198 20.18 0.955 0.05 0.997 4.31 H-32B 1.030 1.279 1-9.45 1.225 15.89 0.968 5.98 1.009 1.99 H-53A 1.109 1.385 19.94 1.325 16.30- 0.960 13.45 1.023 7.78 H-33B 1.o64 1.375 22.59 1.317 19.22 0.954 10.31 1.015 4.63

TABLE II-G (CONT'D) PREDICTION OF VAPORIZATION DATA NG Percent NG Percent NG Percent NG Percent Weir Predicted Deviation Predicted Deviation Predicted Deviation Predicted Deviation Height Liquid NG by Equation Equation by Equatiorns Equations by Equation Equation by Equation Equation System Run No. Inches Rate, gpm Experimental (123) (123) {114)+{124) X114)+(124) (109) (1C9) (108) (108) Freon-12 FIO 87 1-1/2 8.0 2.780 2.818 1.34 2.600 - 6.95 2.665 4.12 2.738 1.49 88 2.780 2.817 1.32 2.598 - 6.99 2.666 4.12 2.739 1.49 89 3.400 3.530 3.69 3.255 - 4.45 3.198 5.93 3.066 9.83 90 3.62 3.50 - 2.53 3.257 -11.16 3.197 11.68 3.065 15.33 91 2.430 2.837 14.36 2.608 - 6.83 2.669 9.85 2.850 17.27 92 2.790 2.837 1.67 2.607 - 7.01 2.669 4.32 2.850 2.16 93 -2.940 2.804 - 4.84 2.579 -13.99 2.620 10.89 2.859 2.76 94 3.250 2.798 -16.12 2.574 -26.25 2.614 19.59 2.850 12.30 He-iC4N90H 147 1-1/2 8.0 1.660 1.525 - 8.82 1.795 7.53 1.308 21.20 1.782 7.35 148 1.520 1.503 - 1.12 1.777 14.48 1.370 9.88 1.781 17.13 149 1.590 1.482 - 7.28 1.756 9.43 1.355 14.80 1.758 10.59 \ 150 1.820 1.528 -19.10 1.787 - 1.84 1.261 30.70 1.781 2.13 151 1.630 1.597 - 2.06 1.874 13.00 1.301 20.20 1.831 12.35 152 1.640 1.683 2.56 1.967 16.62 1.316 19.72 1.901 15.91 153 1.740 1.741 0.06 2.032 14.39 1.351 23.45 1.925 10.61 155 1.510 1.496 - 0.91 1.773 14.82 1.286 14.82 1.739 -15.15 156 1.640 1.467 -11.80 1.741 5.79 1.346 17,96 1.743 6.25 157 1.570 1.462 - 7.38 1.738 9.66 1.343 14.48 1., 34 10.46 N2-iCqH0~ 134 1-1/2 8.0 1.880 1.550 -21.27 1.540 -22.o4 1.184 37.00 1.643 12.59 135 1.800 1.551 -16.09 1.541 -16.80 1.184 34.20 1.644 8.69 136 1.880 1.584 -18.69 1.570 -19.77 1.329 29.30 1.716 8.71 137 1.820 1.552 -17.29 1.-537 -18.40 1.309 28.05 1.691 7.10 138 1.700 1.471 -15.56 1.461 -16.36 1.186 30.25 1.591 6.40 139 1.730 1.471 — 17.61 1.461 -18.43 1.186 31.25 1.591 8.01 140 1.800 1.642 - 9.61 1.636 -10.05 1.169 35.05 1.692 6.02 141 1.840 1.700 -8.23 1.693 - 8.67 1.201 34.75 1.736 5.68

TABLE II-G (CONT' D) PREDICTION OF VAPORIZATION DATA NG Percent NG Percent NG Percent NG Percent Weir Predicted Deviation Predicted Deviation Predicted Deviation Predicted Deviation Height Liquid NG by Equation Equation by Equations Equations -by Equation Equation by Equation Equation System Run No. Inches Rate, gpm Experimental (123) (123) (114)+(124) (114)+(124) 10o) (lo9) (108) (108) He-MIBK 158 1-1/2 8.0 1.740 1.765 1.40 2.154 19.24 1.260 27.55' 1.642 5.65 159 1.750 1.757 0.41 2.152 18.69 1.251 28.50 1.620 7.46 160 1.870 1.856 - 0.73 2.272 17.70 1.515 29.68 1.635 12.60 161 1.870 1.856 - 0.74 2.272.17.68 1.315 29.68 1.634 12.60 162 1.680 1.664 - 0.96 2.0351 17.29 1.201 28.50 1.616 3.78 163 1.680 1.664 - 0.93 2.032 17.31 1.201 28.50 1.617 3.74 164 1.650 1.736 4.98'2.115 21.98 1.239 24.87 1.712 3.78 165 - 1.650 1.731 4.67 2.108 21.72 1.236 25.05 1.706 3.38 He-H20 115 1-1/2 8.0 1.990 1.783 -11.61 1.899 - 4.76 1.796 9.73.1.696 14.78 116 1.920 1.783 7-.69 1.899 - 1.08 1.796 6.44 1.696 11.68 117 1.880 1.750 - 7.45 1.862 - 0.97 1.761.533 1.705 9.30 118 1.850 1.754 - 5.48 1.866 0.85 1.765 4.60 1.709 7.65 10 119 2.140 1.533 -39.61 1.710 -25.16 2.164 1.10 1.937 9.49 120 2.140 1.533 -39.59 1.710 -25.16 2.163 1.10 1.937 9.49 121 2.340 1.586 -47.54 1.771 -32.12 2.465 5.36 2.111 9.80 122 2.370 1.587 049.36 1.772 -33.78 2.465 4.00 2.111 10.92 AIR-I20 64 1-1/2 8.0 2.380 2.021 -17.78 1.568 -51.81 2.478 4.10 2.525 6.10 65 2.420 1.962 -23.34 1.523 -58.87 2.420 0.99 2.468 1.98 -66 3.100 2.138 -44.98 1.654 -87.38 3.055 1.45 2.892 6.55 67 3,180 2.137 -48.84 1.653 -92.39 3.056 3.92 2.892 9.08 68'2.920 2.135 -36.75 1.652 -76.80 3.056 4.67 2.892 9.62 69 2.280 2.118 - 7.64 1.641 -38.95 2.330 2.19 2.483 8.91 70 2.240 2.075 - 7.94 1.608 -39.32 2.294 2.39 2.446 9.19 71 2.220 1.962 -13.17 1.526 -45.45 2.045 7.88 2.245 11.40 72 2.280 1.962 -16.22 1.526 -49.38 2.045 10.30 2.245 1.57 123 2.350 2.116 -11.05 1.651 -42.35 2,319 1.335 2.459 4.62 124 2.340 2.117 -10.52 1.649 -41.89 2.322 0.76 2.465 5.35 125 2.220 2.119 - 4.76 1.652 -34.39 2.238 0.82 2.418 8.90 126 2.110 2.127 0.81 1.660 -27.13 2.242 6.26 2.420 14..67 127 2.290 2.064 -10.95 1.612 -42.03 2.272 0.78 2.409 5.20

-295 - TABLE II-G (CONT' D) PREDICTION OF VAPORIZATION DATA NG Percent NG Percent Ni Percent NG Percent Weir Predicted Deviation Predicted Deviation Predicted Deviation Predicted Deviation Height Liquid NG by Equation Equation by Equations Equations by Equation Equation by Equation Equation System Run No. Inchee Rate, gpm Experimental (113) (123) (114)+(124) (114)+(124) (109) (109) (108) (108) 128 2.440 1.900,-28.39 1.482 -64.63 2.353 3.56 2.398 1.74 129 2.400 1.901 -26.27 1.482 -61.90 2.353 1.96 2.398 0.91 130 2.960 2.038 -45.24 1.587 -86.46 2.943 0.57 2.775 6.24 131 3.140 2.039 -54.02 1.586 -97.92 2.946 6.17 2.780 11.46 53 2 8.o 2.630 2.283 -15.20 1.764 -49.09 2.723 3.54 2.778 5.61 54 2.620 2.322 -12.84 1.794 -46.04 2.758 5.28 2.812 7.35 55 2.570 2.224 -15.54 1.719 -49.47 2.669 3.86 2.724 5.98 56 2.570 2.466 - 4.24 1.917 -34.09 2.432 5.35 2.654 3.26 57 21520 2.483 - 1.48 1.925 -30.91 2.454 2.63 2.681 6.38 58 2.380 2.491 4.47 1.940 -22.70 2.448 2.87 2.668 12 07 59 2.730 2.331 -17.14 1.800 -51.65 2.511 8.02 2.679 1.87 60 2.650 2.351 -12.73 1.816 -45.89 2.517 5.01 2.688 1.44 61 2.480 2.562 3.20 1.978 -25.35 2.524 1.76 2.769 11.63 62 4.010 2.425 -65.36 1.864 -115.07 3.372 15.90 3.200 20.20 63 3.640 2.423 -50.23 1.863 -95.34 3.371 7.39 3.197 12.15 A-32 2 4.58 1.860 2.370 21.54 2.272 18.15 2.269 22.00 2.049 10. 12 A-29 1.700 2.063 17.59 1.974 13.87 1.983 16.65 1.937 13.98 A-19 1.690 1.887 10.45 1.819 7.11 1.815 7.42 1.848 9.33 A-21 1.710 1.645 - 3.95 1.575 - 8.54 1.615 5.55 1.723 0.75 A-31 9.16 1.870 2.622 28.68 2.512 25.57 2.459 31.50 2.183 16.75 A-28 1.860 2.180 14.77 2.086 10.82 2.067 11.13 1.999 7.50 A-18 1.760 2.103 16.32 2.017 12,76 1.982 12.62 1.996 13.43 A-20 1.640 1.911 14.19 1.826 10.16 1.818 10.81 1.929 17.60 A-30 18.31 2,170 3.118 30.40 2.977 27.11 2.861 28.55 2.450 12.91 A-22 2.110 2.656 20.56 2.537 16.81 2.422 15.29 2.304 9.19 A-23 1.970 2.593 24.03 2.471 20.27 2.340 18,78 2.332 18.33 A-24 1.800 2.406 25.19 2.291 21,42 2.170 20.55 2.270 26.05

-296,TABLE II-G (CO1NT'D) PREDICTION OF VAPORIZATION DATA NG Percent NG Percent NG Percent NG Percent Weir Predicted Deviation Predicted Deviation Predicted Deviation Predicted Deviation Height Liquid NG by Equation Equation by Equations Equatiors by Equation Equation by Equation Equation System Run No. Inches Rate, gpm Experimental (123) (123) (114)+(124) (114)+(124) (109) (109) (108) ((108) A-26 32.04 2.550 3.794 32.78 3.609 29.35 3.312 29.85 2.884 13.98 A-25 2.580 3.333 22.59 3.166 18.50 2.893 12.11 2.727 5.69 A-27 2.560 3.507 26.99 3.327 235.05 2.950 15.23 2.935 14.62 AD-39 4.58 1.960 2.342 16.31 2,256 15.10 2.211 12.82 2.030 3.57 AD-40 1,610 1.844 12.70 1.754 8.20 1.788 11.09 1.839 14.24 AD-45 1.720 1.625 - 5.81 1.540 -11.67 1.601 6.90 1.711 0.51 AD-58 9.16 2.050 2.539 19.26 2.441 16,01 2.357 14.97 2.153 5.05 AD-37 1.780 2.027 12.18 1.930 7.76 1.921 7.90 1.975 10.96 AD-44 1.580 1.641 3.75 1.553 - 1.71 1.620 2.50 1.748 10.60 AD-54 18,31 2.720 2,968 8,55 2.829 3.85 2.681 1.44 2.450 9.95 AD,33 2.530 2.667 5.16 2.535 0.21 2.409 4.78 2.360 6.71 AD-43 2.?330 2647 11.99 2 526 7.75 2.390 2.56 2.327 0.12 AD-36 2.490 2.497 0.29 2.372 -4.98 2.259 9.26 2.303 7.51 AD-41 32.0 2.810 3.626 22.50 3.458 18.74 3.125 11.20 2,838 1.00 AD-42 2.820 3.523 19.96 3.349 15.81 2.981 5.70 2,897 2.72 A-2 5-1/2 4.58 2.13P 2.699 21.07 2.547 16.36 2.550 19.72 2.285 7.29 A-9 2.340 2.523 7.27 2.425 53.43 2.322 0.78 2.249 3.88 A-li 2.470 2.415 - 2.26 2.332 - 5.92 2.190 11.34 2.227 9.84 A-17 2.550 2.400 - 6.25 2.321 - 9.86 2,174 14.72 2.206 13.50 A-13 2.390 2.332 - 2.47 2.226 - 7.35 2.112 11.60 2.251 5.85 A,3 9.16 2.280 3.020 24.51 2.873 20.64 2.753 20.76 2,467 8,22 A-8 2.640 2.818 6.32 2.702 2.31 2.524 4.40 2.413 8.60 A-10 2.760 2.729 -1.14 2.611 - 5.72 2.425 12.12 2.447 11.31 A-16 2.740 2.710 - 1.09 2.598 - 5,46 2.401 12.38 2,427 11.41 A-12 2.510 2.700 7.06 2.576 2.57 2.367 5.70 2.494 0.62 A-4 18.531 2.730 3.358 18.70 3.192 14,47 2.992 9.60 2.645 3.13 A-5 2,620 3.378 22.43 3.218 18.59 2.997 14.40 2.658 1.45 A-7 3.270 3.291 0.64 3.140 4.15 2.857 12.62 2.699 17.47 A-15 3.100 3.289 5.74 3.137 1.17 2.854 7.94 2.699 12.92 A-14 18.31 3.190 3.422 6.77 3.265 2.29 2.888 9.46 2.867 10,11 A-6 32.0 4.450 4.429 - 0.47 4.191 - 6.17 3.715 16.52 3.304 25.75 Absolute Average Percent Deviation 14.59 21.08 19.62 11.59 Standard Deviation 0.389 0.502

TABLE III-G PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN, CABBON DIOXIDE-CYCLOHEXANOL SYSTEM WEIR HEIGET, 3-1/2 IN; SPLASH BAFFLE HEIGHt, 4 IN. RUN NO. 2-A 2-B 3-A 3-B 4-A 4-B 21-A 21-B 5-A 5-B 19-A 19-B 8-A 8,B BAROMETRIC PRESSURE, IN Hg 29.13 29.12 29.19 29.19 28.96 28.96 28.77 28.77 29.04 29.'04 29.30 29.30 28.92 28.92 LIQUID FLOW ROTAMETER READING 15.0 15.0 15.0 15.0 15.0 15.0 15.0 15.0 15.0 15.O 15.O 15.0 15.O 15.0 LIQUID TEMP. AT ROTAMETER, ~F 102.1 10~.6 99.7 99.7 99.4 99.9 97.0 97.6 99.9 99.6 100.5 I01.4 76.0 76.1 LB. MOL PER MINUTE 0.3840 0.3840 0.3840 0.3840 0.3840 0.3840 0.3849 0.3848 0.3840 0.3840 0.3840 0.3840 0.5860 0.5860 GALLONS PER MINUTE ~. 950 ~. 950 4.950 4.950 4.950 ~. 950 4.937 4.93 ~. 950 4.950 4.950 ~. 950.... GALLONS PER MIN. PER FT. WEIR 7.920 7.920 7.920 7.g20 7.~720 7.920 7.899 7.899 7.920 7.g20 7.g20 7.g20.... GAS FLOW AT ROT4t~TER ~OTANETER READING 120 120 160 160 40. O 40.0 80.0 80.0 80.0 80.0 162 162 ~3.0 kO. 0 TEMPERATURE AT ROTANETER, ~F 105.l 104.0 105.9 105.9 85..1 87.5 91.8 9~.5 91.~ 93-5 112.0 112.O 80.0 81.0 PEESSURE AT ROTAMETER, IN Hg 30.73 30.52 32.19 32.t9 31.16 31.36 30.27 30.},9 31.39 31.44 31.80 31.80 30.8~ ~0.72 LB. M0L PER MINUTE 0.2362 0.2400 0.3220 -- 0.0865 -- 0.1575 O.1570 0.1610 0.1621 O.3220 0.~220 0.0864 0.0809 CUBIC FEET PER MINUTE 95.0 96.7 123.7...... 62.6.0 62.66 62.0 62.30 126.9 126.9 35.04 50.60 GAS FLOW AT TEST TRAY SUPERFICIAL GAS VELOCITY, Vs, FT/SEC. 2.680 2.8~ 3.366 3.256 0.8910 -- 1.750 1.743 1'730 1.699 3.670 3,480 O.8910 0.8570 DENSITY, PG' ~B. PER CU.FT. 0.1007 0.1014 0.1092 0.1092 0.1089 0.1097 0.1058 0.1062 0.1089 0.1079 0.1080 0.1077 O.11~6 0.1155 I F-FACTOR, Vs4PG 0.859 0.901 1.1i2 1.112 0.294~ -- 0.5695 O. 568a 0.5710 0.558 1.207 1.141 0.5004 O.2geO 53 kO HYDRAULIC DATA~ -— ] FROTH HEIGHT, IN. 7.6 7.6 8.4 8.4 5.5 5-5.... 6.7 6.7 8.6 8.6 6.0 6.0 i CLEAR LIQUID HEIGEr, IN POSITION 2 4.30 3.40 3.50 5.50 3.40 5.34 3.75 3.75 3.75 3.80 5.30 5.50 5.80 5.80 POSITION 5 5.20 5.30 2.50 2'..50 3.40 5.0 5.15 5.15 ~.10 5.10 2.45 2.45 5-55 5-55 POSITTON 4.... 2.40 2.40 5.20 5.~0 5.25 5.25 5.50 5.50 2.40 2.~0 5.50 5.50 POSITION 5 5.0 5- 0 5- 50 5.50 5.40 5.37 5.70 5.70 5.66 5.56 5.20 5-20 5- 70 3.70 PRESSURE DROP ACROSS TRAY, IN H20.... 4.60 4.60 2.90 2.95 5.50 5.50 5.20 5.50 4.80 4.95 5.10 2.95 GAS COMPOSITION, MOL FR Y (ENTERING TRAY).... O. 9131 -- O. 9110 O. 9159 -- O. 911~ 0.8764........ Y(LEAVING TRAY) 0.8294 0.8293 0.9114 0.9114 O. 9139 0.9148;'9521 0.9407 0.9046 0.8702 0.9096 0.9084 0.9428 0.9451 LB.MOL LIQUID COMPOSITION, - FT~- X 105 * POSITION 1 25.37 24.59 26.7 25.]'8 27.12 25.36 19.1~ 19.72 28.93 25.1~ 18.60 18.05 24.90 25.20 POSITION 2 107.83 114.17 126.12 126.29 85.95 73.44 98, 97 100.90 124.22 t11.42 117.90 115.5 55.80 56.75 POSITION ~ 104.77 116.41 138.04 1~6.18 98.79 104.91 112.4 114.50 124.90 115.O4 129.80 126.9 8~.40 81.80 POSITION 4 -- 140.58 159.28 174.48 112.28 114.09 12~.8 121.06 130.96 124.59..... 87,00 84.36 POSITION 5 148.03 149.~ I72.54 128.4~ 126.12 154.0 142.50 152.57 156.45 157.0 156.0 100,6 106.1._ POSITION 6 t5~. 55 t54.9~ 151.87 180.05 131.8~ lI~. 50 148.2 149.60 128.27 149.50 164.9 165.t 113.0 114.0 EQUILIBRIUMS CONDITIONS ON TEST TRAY PRESSURE ABOVE TRAY, AT~ 1.O0~7 0.9901 1.0~9 1.O59 1.O35 t.041 0.9967 O.998~ 1.03~6 1.0380 1.Ce80 1.0~60 1.O~20 1.0180 TE~PERATURE, ~F 100.2 ~00.2 99.72 99.72 99.41 99.86 99.68 100.15 99.86 99-59 99.64 100.3 72.64 72.75 CO2 PARTIAL PRESSURE, ATM 0.8324 O.8211 0.9469 0.9469 0.9459 0.9525 O.9~52 0.9422 0.9354 0.9035 0.9~51 0.9320 0..9635 0.9621 HERU~Y'S LAW CONSTANT, ATM/MOL FR 257.0 2'57.0 256.0 256.0 255.5 256.4 256.0 256.8 256.4 255.9 255.9 2~7.4 222.5 222.5 x..?, M0L FR X 105 ~2~.0 319.5 36'9.8 369.8 370.2 ~71.4 565~ 566.9 56~.8 555.7 565.4 362..1 ~5.~ ~52.8 C*, LB.NOL PER CU.FT. X lO5 188.5 185.8 214.9 214.9 215.1 215.9 212.4 213.5 212.1 205.3 212.5 210.9 258.7 257.9 MUEPHREE EFFICIENCY % Eml 16 78~60 80.69 66.28 80.40 55 ~40 57~23 67.18 67~30 54~!O 68~66 75.~!8 79.~6 ~8.20 38,62 Eml 15 75 O0 77.50 77~15 53~65 )~ ~, ~'~"1.~i. 7i.09 71~5~-'~.....? _ ~, --,.,.....~:,~.. o ~ 5;5 ~, iS ~ ALiL S~'J;.~PLES T-/%i{EN BY USE OF I{.,'.'IODEPt~.(IC SYi~iING-E

TABLE III-G (CONTINUED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSITY OF MICHIGAN, CARBON DIOXIIE-CYCLOHEXANOL SYSTEM WEIR HEIGHT, 3-1/2 IN SPLASH BAFFlE HEIGHT, 4 IN. RUN 30. 7-A 7-B 6-A 6-B 9-A 9-B 16-A 16-B 20-A 20-B 14-A 14 -B 12-A 12-B BAROMETRIC PRESSUREJ, IN Hg 29.07 29.07 29.40 29.40 28.92 28.92 29.33 29.33 28.77 28.77 29.31 29.51 29.10 29.10 LIQUID FLOW ROTA ETER READING 15.0 15.0 15.0 15.0 15.0 15.0 15.0 15.0 15.0 15.0 50.0 50.0 50.0 50.0 LIQUID TEMP. AT ROTANETER,"F 78.4 79.4 77.7 W7.0 81.6 81.9 80.6 81.1 75.0 75.3 77.40 76.40 74.6 75.9 LB. MOL PER MINUTE 0.3860 0.3860 0.3860 0.3860 J.3860 0.3860 0.3860 0.3860 0.3865 0.3864 1.294 1.292 1.295 1.295 GALLONS PER MINUT 4._92 4.920 4.920 4.920 4.92o 4.920 4.920 4.920 4.910 4.910 16.370 16.490 1i6.450 16.460 GALLONS PER MIS. PER FT. WEIR 7.872 7.872 7.872 7.872 7.872 7.872 7.872 7.872 7.856 7.856 26.19 26.38 26.32 26.34 GAS FLOW AT ROTAMETER ROTAMETER HEADING 80.0 80.0 80.0 8o.o 112 112 160 160 162 162 41.o 41.0 80.0 8i.o -TEMPERATURE AT ROTAMETER, F 80.6 81.7 81.0 81.6 85.8 86.6 95.4 96.2 92.0 93.8 84.5 84.5 78.8 80.75 PRESSURE AT ROTAMETER, IN Hg 30.97 30.97 31.20 31.10 31.32 31.32 31.73 31.78 31.32 30.51 30.51 30.70 30.70 LB. MOL PER MINUTE 0.1610 0.1600 0.1615 0.1616 0.2250 0.2260 0.3220 0.3220 0.3250 0.3239 0.0821 0.0822 0.1621 0.1638 CUBIC FEET PER MINUTE 61.50 61.60 61.20 61.60 85.75 86.30 123.6 123.3 125.2 1325.8 31.99 32.00 62.08 62.94 GAS FLOW AT TEST TRAY SUPERFICIAL GAS VELOCITY, vs, FT/SEC. 1.670 1.673 1.675 1.676 2.352 2.364 3.340 3.341 3.428 3.431 0.8631'0.8619 1.707 1.730 DENSITY, pG, LB. PER CU.FT. 0.1128 0.1122 0.1129 0.1121 0.1123 0.1111 0.1122 0.1119 0.1109 0.1103 0.1103 0.11o6 0.1095 0.1093 F-FACTOR, VW[PG 0.561 0.561 0.564 0.562 0.7890 0.790 1.120 1.119 1.142 1. 139 0.2866 0.2867 0.5688 0.572 HYDRAULIC'DATA.C FROTH HEIGHT, IN. 6.6 6.6 6.8 6.8 7.4 7.4 8.6 8.6 8.7 8.7 741 7.7 9.6 9.6 CLEAR LIQUID HEIGHT, IN POSITION 2 3.70 3.70 3.60 3.60 3.80 3.80 3.35 3.35 3.40 3.40 4.6o 4.60 5.20 5.20 POSITION 3 3.05 3.05 3.20 3.20 2.95 2.95 2.50 2.50 2.55 2.55 4.35 4.35 4.20 4.20 POSITION 4 3.25 3.25 3.30 3.30 3.00 3.00 2.50 2.50 2.80 2.80 4.50 4.5o 4.40 4.40 POSITION 5 3.75 3.75 3.80 3.80 3.60 3.60 3.30 3.30 3.60 3.60 4.90 4.90 5.00 5.00 PRESSURE DROP ACROSS TRAY, IN -H2 -- -- 6.6 4.o 4.o 5.0 5.0 5.0 5.10 4.0 4.1 4.6 4.7 GAS COMPOSITION, MOL FR Y(ENTERING TRAY) — - -- --- --- ---------- Y(LEAVING -TRAY) LB.MOL X 0.9378 0.9474 0.9294 0.9211 0.9395 0.9200 0.9430 0.9144 o.9263 0.9266 0.9056 0.9056 0.8911 0.82 LIQUID COMPOSITION, LB.MOL X 10F * POSITION 1 28.38 28.91 26.84 27.10 30.42 32.14 28.45 26.10 29.95 29.39 32.60 34.55 43.45 46.10 POSITION 2 -- -- -- -- 109.8 116.o 85.85 125.7 123.8 121.4 34.33 35.18 53.45 53.90 POSITION 3 104.31 104.8 90.0 124.o 119.1 127.6 127.8 128.4 131.4 33.83 38.20 62.70 63.60 POSITION-4 -- -- -- - 137.7 139.9 -- -- -- -- 59.00 65.50 87.50 86.50 POSITION 5 132.9 132.9 126.2 -- 145.8 145.5 154.7 152.9 151.3 151.1 67.35 -- 96.10 97.05 POSITION 6 138.8 141.8 137.8 137.5 153.7 158.1 164.3 163.7 161.7 160.6 74.00 81.95 101.8 104.6 EQUILIBRIUMS CONDITIONS ON TEST TRAY PRESSURE ABOVE TRAY, ATM 1.0220 1.0150 1.c.6o 1.0240 1.0220 1.0217 1.o26o 1.0254 1.0084 1.0033 1.0080 1.0130 1.0090 1.0077 TEMPERATURE, "F 75.65 76.01 76.3 76.46 79.16 80.13 77.43 77.99 76.64 77.09 76.91 76.82 77.18 77.76 CO2 PARTIAL PRESSURE, ATM 0.9584 0.9616 0.9536 0.9432 0.9602 0.9400 0.9710 0.9376 0.9341 0.9297 0.9129 0.9174 0.8991 0.9011 HENRY'S LAW CONSTANT, ATM/MOL FR 224.5 224.7 225.0 225.2 228.2 228.8 226.1 226.7 225.3 225.8 225.6 225.5 225.8 226.4 x*, MOL FR X 105 426.9 427.9 423.8 418.8 420.8 410.9 429.0 413.6 414.6 411.7 404.6 406.8 398.2 398.0 C*, LB MOL PER CU.FT. X 105 252.8 253.0 250.5 247.6 247.4 241.1 252.5 243.7 242.1 243.1 238.8 240.2 234.8 234.6 mrmPHEEE EFFICIENCY % Emi 16 49.51 50.62 50.05 50.54 56.83 60.16 60.94 63.59 61.74, 63.82 17.33 22.35 30.65 31.15 Eml 15 46.90 46.63 44.71 -- 53.17 54.15 56.41 59 2' 56.19 57.08 165 87 -- 27.63 27.02 *ALL SAMPLES TAKEN BY USE OF HYPODERMIC SYRINGE

TABLE ZITI-G (CONTVMED) PLATE EFFICIENCIES IN 2ECTANGUIAR -COLUMN AT UNIVERSITY OF MICHIGAN, CARBON DIOXIDE-CYCLOHEXANOL SYSTEM WEIR HEIGHT, 3-1/2 IN; SPLASH BAFFLE HEIGHT, 4 IN. RUN NO. 10-A 10-B 17-A 17-B 15-A 15-B 13-A 13-B 11-A 11-B 18-A tBAROMETRIC PRESSURE, IN Hg 29.05 29.05 29.31 29.31 29.18 29.18. 29.10 29.10 24.O5 29.05 29.31 LZIoUI FLOW ROAMSTER ftADING 50.0 50.0 50.0 50.0 82.0 84.,o 80.0 80.oo.0 8o.o 80.0 8 0.0 LIQUID TEMP. AT ROTAMETER, ~F 77.5 77.8 77. O 77.0O 76.0O 77.3 80.1 80.0O 81.3 82.0 77.3 7. LB. MOL PER MINUTE 1.292 1,294 1.294 1.294 2.117 2.16-9 2.0O68 2.068 2.069 2.O068 2.070 2.7 GALLONS PER MINUTE, 16.4-90 16.460 16,455 16.455 26.904 27.583 26.330 26.330 26.36 26.36 26.33 2. GAL LONS PER MIN. PER FT. WEIR 26.38 26.34 26.33 26.33 43-.05 4~. 13 42.13 42.13 42.18 -42.18 42.13 4.1 GAS FLOW AT ROTAMETER ROTAMETER READNG- 108 108 158 158 41.Jo 41.0 80. O 78.0 107 107 156 1 5 TEMPERATURE AT ROTAMETER, ~F 82.4 83.1l 95.8 96.8 85.1 86.9 85.4 85.6 75.9 86.3 98.3 8. PRESSURE AT ROTANETER, IN Hg 31.O5 30.95 31.68351 05 }1.35 30.65368 13 LB. MOL PER MINUTE 0.220 0.220 0.3233 0.3235 0.0O8244 0.08215 0.1-620 0.1576 0.2187 0.2171 0.3226 052 CUBIC FEET PER MINUTE 83.8 -84.2 124.4 125.1l 31.95 32.27 63.22 6-1-B3 83.43 84.4212. GAS FLOW AT TEST TRAY -SUPERFICIAL GAS VLCTY, sFT/SEC. 2.300 2.315 3.387 3.390 0.8638 0.S710 1.742 1.695 2.342 2.331 3.4o~ 542 DENSITY, PG, LB. PE CU.F'T. 0.l098 O.1085 0.1074 0.1067 0.110~4 0.1085 0.1054 0.1050 0.1'063 0.10_55 0.1.033.le F-FACTOR, ~FpG 0.7620 0.7626 1.110 1.108 0.2870 0.2869 0.5656 0.54-93 0.7,635 0.7571 1.093.0~ HTORtrr.C DATA k FROrH HEIG~r, IN. 10.5 10.5.10.3 10.7 8.8 8.8 10.8 10.8 11.5 11.5 12.1.0 CLEAR LIQUID HEIGHT, IN POSITION 2 4.055.95 4.95 6.-ol 6.O1 6.20 6.20 6.05 6.05 5.9 59 POSITION 3 4.15 4.15 3.70 3,70 5,40 5.-4O 5.10 5.3-0 4. 80 4.86 4.55.5 POSITION 4 4.25 4.25 3.65 3.65 5.50 5.50 5.2,0 5.20 4.80 4.80 4-.45 44 POSITION 5 5-.00 5.-OO 4.40 4.40 5.-90 5.90 5.90 5.90 5.55 5.55 5.10 51 PRESSURE DROP ACROSS TRAY, IN N20 5.2 5.2 6.1 6.0.... - 5.5 5.5 5.80 -5.95 7'.20 72 GAS COMPOSITION, MOL FR Y (ENTEiNG TRAY)............. Y (LEAVING TRAY) 0.8699 O'.8623 0. 8229 0.8059 0.-9005 0.8827 0.8389 0.8356 0.8512 0.8389 0.7352 070 LIUTD COMP0SITION LB.MOL X]0 POS$ITION 1 46.80 5]..50.54.80 52.50 41.0 38.10 43.00 42.78 49.04 25.98 52.30 5.f POSITI'ON 2 63.-O0 48.00 85.4O 80.40 -41-45 39.00 47.22 47.38 54.90 -- 66.4O 6.O POSITION 3 7 1.70 75.50 88.,O0 84.90 41.00 39.40 55.00 53.50 60.90 63.90 68.30700 POSITION 4'92.-50 90.90 111:. 110.1 63. 50 62.90 70.10 69.80 77.25 75.50 95.10 9.0 POSITION 5 104.0o 10~.6 111.3 110.4 65.50 -6T.30 8i.40 83-. 60 8.O 90.50 96.60o 2.O POSITION 6 109.3 1-13.0 119.0 119.2 74-.60 73.O0 90.09 90.10 93.00 94.10 98.6010. EQUILIBRIUMS CONDITIONS ON TEST TRAY RESSUR ABOVVE TRAY, ATM -1.0180 1.0110 1.0130 1.0130 1.0120 1.0020 0.99T7 0.9977 1.O01 0.9993 1.O080 106 TEMPERATURE, ~F 77.0O 77.63 76.28 76.$6 75.20 7,6.87 82.13 82.22 81.68 82.26 77. 187.5 C02 PARTIAL PRESSURE, ATM 0.8856 0.8718 0.8336 0.-81-64 0.9113 0.w8845 0.8370 0.8337 0.8521 0.8383 0.7411 075 HENRY,'S IAW CONSTANT, ATM/MOL FR 225.6 226.3 225.0 225.2 224.0 225.2 231.0 231.0 230.5 231.1 225-.8 2 6. xMOL FRX15392.6 385.2 37-0.5 362:5 4_06.8 382.1'362.3 360.9 369.7 362.7 328.2 2. C *, LB. MOL PER CU.FT. X 105 232.2 2.27.0 218.9 214.2 241.1 231.5 212.1 211.1 -216.5 212.2 193.519. MJR-PH1REE EFFICIENCY % Eml 16 33 ~99 35.14 39. 38 41.54 -17.O0 16.57 27.71 28.-07 26.11 36.44 32. —85 3.9 Eml 15 30.93 30.26 34.54 35.91 12.31 13.91 22.66 22.86 23.21 34.46 31.-428.6 ~ALL SAMPLES TAKEN BY USE OF HYPOIERMIC SYRINGE

C!T {"T (BO Di OK! D%. CYO i0F;/NOhS;'~;J'{'}EiR~ J!~Gl!HE 2 IN.; SF]%~ 3-l!:,171:!~, 2-/2.L!', NTI~~~~~~~~~~~j NO, 26-A ~~~~~~~~~~~~~~~~~~~~~~~~~~24-A 23-2 23- 25-B 49-A 7PTjlj~~~~~~~~~~~~~~i NO.:)6 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~2-A~ 26-.AQ-~ 23- 34-B 25-A 2 - 9 A 4BAIR.ONI~~IRIC!P, ES'SURE', IN PH' 29.08 29.08 29o12I 29, 1 2 29153 29.53 2'::4.0 29.4-0 29.09 29109 29.46 2'9. 46 29.382.3 LIQUID FLOW ROTAPTE2'7R READIN',G 15.0 15.0 15.0 15.0 1 -5.0 1-5.0O 15~ 0 1-5.0.5~ 0 15.0 15,0 15.0 15.0 LiP-iE)i TEMTP., AT FOTAMETER,~F 1 00,3 100.2 93,'6 95.1 ]_0!.6 i00. 7 98".5' 99.5 1!. 101.8 99.6' 99.1 99.8 00. LB. MOL PER?0Z,rO' 385 0- 38 5 C0) 385 O. 38~ 0 387 O.- i-~,385 0.38411 0.38,,n O 385 O. 385 0. 385 O. 38-5 O GAIZZ)NS.P~~~~~~~~~iR P[!}Gj~~~~~~~i!.C ~~~4.9'f32 4 932 4. 932 4. 934- 4. 936 4,..94 1 4.9- 931 4.93 9194 GALLONS.PER MI~,~ P7ER. FT. WFEIR 7.891 7,.891 7.891 7.894 7.898 7,9O 6 o 7.890 ~ )" 7.89?.906 7.906 7,,891 7.,891 7.0 GAl'S FLe, T AT RTIdTER 46 4. 5 83 83~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~8 183 6.!MtE }l,'\FueE AT- PT'ANS'.ER,'?F 83.683, 6 9,!'. i0. l 6 3L8 165 14. 147.3 103, 7 1 0L415.4. PN~E SST5,'i~S AT TER'~',~R IN Pig 31,23 3.8 0,7 30. 72- 30.7'7`6{', 316 3i.-96 -3,6316 31o79 31 2.98 B. -B L P l I1 N, Nii}7JT- E, -. O.!0!5, 102 1 D. 16'36 O. 16:52 0.2363 0.35,339026,2;34 0.28 30 28930 0.2894 Cb-slC rTET PFR M!KL~~~~~~~~~~fE ~38.56 374Y40 63,99 93-90 9 i 4 1.65 -123.0 11. 118.67- 111.31142 GAS FLOW1 AT'.FEST TPJff SU.FRF!!\ Gi. T1,$-~ VE DOC!TY, vs F7 /SEC. 1, 0i. 0 9 1,78 1.783-' 2.548 2_.,5416 309'~2 3.1 7'986 2,988 3 0 6 3 1 0 422 1~0 Z0Ei-TSiT-Yr p:'f~!, PR CU~ FT. 0,'1116 0,1118 02109-1 0.10og 0.110-8 0'. I1iO; 0.0.9334 0.0882 0.1019 0.-10-17 0.11160.0! 0.1006l O. 1 0 04o F-,FACT~~~~~~~~~~~ ~ ~ ~~~~~~~~~~~~~~OR.v~ ~,3607 0,3 643 0. 5909 O. 5884 O. 846.5 O~ 8.!;83 0., 9443 O ~7 O 952 9 O.930 i.04O 90.362 i 6 P!D~,J!CDATA... F P, 0TH i-,!GHT,!N, 4.0O 4, 1 4.7 4.7 5.4 57 6"0 6.0o. 5.4 5.8 5.8 8.0 8. CIEL!LR LIQUID!HE!GHT,. IN POSITION 2 2.2 0 2.20 2,!0n 2.10 2.10 2-!0... 2, I0 2.10 2.00 2.00 2.00 ",0 POSITION 3 2, 00, 2.00 1.80 1.8o 1. 60 1.60... 1.6r0 1.60 1.45 1.45 1,45 POSITION 4 2.07 2.07 2.00 2.00 1,6) 1.60O -L...'i60 i. 60 1. 40.1.40' 1,40 I'!0 POSITION 5 2. i0 2.10 2.1!0 2.10 2.!0 2.i0,.... 2.1t0 2.1i0 2. O)'- 2. 05 2.05 2.0 P7RESS'iLrF~E DIRECT 2CROSS TRAY, IN H20 2. 0 1.90 2.30 2.30 2.90 2,0. 5 2.95 }O 1.! 3.43 340.546,GAS COHPOSITE11 —iON~1 0L..' FR y(LEA"VIPG TA)0.9779 0.9766 0.9742 0.9755 0.9 768 0.97-50 0.:50-17 0 50!7 0.697-5 0.971O642% LIWD OIIPOITOI1,LB.IMOLX 0 POSITION 1 23.94 24.22 26.12 10.60 27.56 27.-59 I15. 11 14.Oc2! }.24 18. 19 2 5.9 2 25.45 17.52169 POSITION 2 33.09 92.14' 30.O1 31.47 54.37 69.46 79.96 81.,31 754.51 34 7 475075 1',65 88 63.446.0 POSITION 3 61.18.-102.9 64.23 57,46 82.90o 83.74 98.77 lo0. 49 42.72 43.80 60.-,07 69.56 102.7 105.92769 POSITION 4 76.24 64.59 91.62 89.56 119.93 120.55 44.34- 50. t0 71~ 75 89.29 133.0 132.2993 POSITION 5 97.36 132. 83 140.0o. 97.14 95.9!i3.64 11.2 13.6i.9 52.25 53.58 8.4 95 -99 i133,6 140. 64 L01.4 0.9 POSITICION 6 109.4 102 2.6 -1258 419 ~4.~ 60.. 9 5 8 -.7 9234 1,01.03 i l 9.6 0 149.6 E"JLBRUSCON'ZDITIONS ON TEST TRAY PlSSSUPJ-H ABC',-E- TRAY, ATM 1.0354 1.0371 1.0134 1.0150 1.0287 1 L.02 94 1.0414 1.04-14 1.0525 1.-0515 1.0348 1.0314 1'6 TEMFFEPtTb~E, "'F 99-9 99.9 100.2 100.5 100.4 919.7 100.9 101.7 o00. 4 100.5! 00. I 100.2 LO0.410. CO,,, P?,~TIAL PRESSURE, ATM 1 012PO 1.01,27 0.9872 0.9900 1.0055 1.0016 0.5224 0.5225 0.7199 0.7174 1..0050 1.0026 067 HI3E-YYS LA.WCOvT' /T, ATM/MOL FR 256.3 256.5 256.8 257.6 257.3 256.1 2'5, 27.2 257.257.. 257:'7.3 276 X*; MIOL FE~ X l0-? 394.8 394.8 384.4 384.3 390.'8 391.! 203_.0.202.6 27Q.~ 278.6 391.7 390.1 -?63.426. C *, LB. M0LI PER CU. FT. X 105 229.6 229.5 223.6 223.51 227.2 227-3 11i7.6 1 17.4 162.-7 161.9 227.2 22 6. 8 L53.212. M1LURPH,%RE E EFFi'C —-N',C Y % -E-m ra -6~"59 41.76 50.24 53.97 57.:16 07~ o 42 $! 5.47 57.52 61.68 61..46 64.2167 7 Eml -15 56 4 44.23 48.54 5.6 5.a 3.02 38. 04 47 14 ".63 56.67616 ~S A.MPLES W',ITTIERAWN BY USE OF HYPODERMIC SYRINGE, OTHER Sf1{PiiS WITHDRAWN THROUG~H VIALVnE!'NTO S.{:w OTTLE-r

TABLE III-G -(CONTINUED) PLATE EFFICIENCIES IN RECTANGULAR COLUMN AT UNIVERSlITY OF MICHIGAN, CARBON DIOXIDE-CYCLOHEXANOL SYSTEM WEIR HEIGHT,.2 IN; SPLASH BAFFLE HEIGHT, 2-1/2 IN. RUN NO. 45-A 45-B 29-A 29-B ~2A 42-B 32-A 32-B 41-A 41-B 4-A 7 -B3- A 5B BAROMETRIC PRESSURE. IN Hg 29.12 29.12 29.26 29.26. 29.27 29.27 29.06 29.06 2 9.182912.02.O2.2 2.0 LIQUID FLOW ROTP14ETER READING 15. O 15.0 15.~0 15.~0 15.~0 15.0 15.O 15.~0 15.0 15.~0 50.0 50.0 50.0 50 LIQUID TEMP. AT ROTAMETER,:F 100.8 100.8 l1.O- 101.7 101.9g 101.7 1 00.6 102.1 99.9 1002 002 00390 LB. M0L PER MINUTE 0.384 0.385 0.385 0.385 0.385 0.385 0.385 0.385 0.385 0.385 1.285 1.285 1.288 1.8 GALLONS PER MINUTE 4.940 4.930 4.929 4.929 4.940 4.946 4.946 4.941 -4.94o 4.940 16.529 16.530 16.'540 1.4 GALLONS'PER MIN. PER FT. WEIR 7.904 7.888 7.886 7.886 7.904 7.914 7.914 7.906 7.190% 7.904 26.446 26.448 26.464 2.6 GAS FLOW AT ROTAMETER ROTAMETER READING 1 92 190 2 03 2 05 2 53 2 53 300 500 313 513 50.0 50. O 65 6 TEMPERATURE AT ROTAMETER, -F 133.0 133.0 134.5 135.5 147.0 147.2 16i.0 161.3 t-61.2 161.o 131.3 131. 6 130.7 PRESSURE-AT ROTAMETER, IN Hg 31.62 31.57 31.56 31.61 32.67 32.69 33.16 33.16 30.68 30.68 50.-92 30.-82 31.16 LB. MOL PER MINUTE 0.4045 0.4017 0.4419 0.4464 0.5183 0.5197 0.6211 0.6294 o.6386 0o.6384 026.00027.CUBIC FEET PER MINUTE i65.5 164.7 i:81.6 t183.5 210.1 210.6 253.8 257.3 -257.0 256.9 86.0o 84.6o950 -GAS FLOW AT-TEST TRAY SUPERFICIAL GAS VELOCITY, vs FT/SEC. 4.431 4.408 4.905 4.,959 5.632 5.649- 6.809 6.902 6.-14 6.9135.4.6 DENSITY, pG'-PER CU. FT. mo. 4 0. 1052 0.0923 o.o09265 0. 09850 0. 09849 0092 091 0059 oogo oo6.io.lop F-FACTORI v 1.4313~~~~~~ 1419 -.49Ot 1.505 1.7690 1.7738 2. o460 2..071 2-143 2.143 0.735 O. 7140 0.780 HYDRAULIC DATA FROTH ~HE IGHT —I'N. 8.5 8.5 8.0 O 8.0 100 0.. 00 1.870 7077 CLEAR LIQUID HEIGHT, IN POSITION 2 2.0OO 2.0OO 2.0O0 2.0O0 2.20 2.20 2.15 2.15 2.15 2.15 3.55 5.55 3.55 m POSITION 3 i. 40 1.40 1.50 1.50 1.70 1.70 1.80 1.80. 1.85 1.85 2.70 2.70 2.65 26 POSITION 4 1.35 -1.35 1.40 1.40 -1. 60 1.6o 1.70 1.70 1.75 1.75. 82802.75 27 POSITION 5 2.00 2.00 2.10~ 2.10 2.20 2.20 2.20.2.20 2.15 2.15. 3.30 5.50 3.25 52 PRESSURE DROP ACROSS TRAY, IN H20 5.40 5~ 55 5.50 5.,50 6.30 6.2............... GAS COMPOSITION, MOL.FR y -(ENTERING TRAY).............. ~y(LEAVING TRAY) 0.8416 0.8249 0.5645 0.5645 0.6889 0.6890 0.4783.71057.97089.14 0 7 ~068 LIQUID COMPOSITION, LB. MOL X ljO5 FT) POSITION i 18.45 16.49 16.06 22.22 14.20 -- 3.49 4..50 15.4 13.6 27.48 26.40 26.03 42 POSITION 2 76.90 77.91 68.66 4-0.36 69-.68 66.33 48.67 50.7-0 7D. 92 69.11 35.01 34.44 34.22 5.9 POSITION 3 1ot. 33 87.45 69.74 86.24 93.68 104.77 68.44 50.26 102.14 95.71 38.8o 41.27.35.59 562 POSITION 4, 130.30 129.36 97.6o 106.79 119.19 120.87 84.08 1.8 116. o1 115.38 64.9o 64.34 57.255.9 POSITION 5 125.52 120.46 93.70 -- 110. 92 lll.-20 76.64 77.33 105.71 105.52 84.25 81.00 70.187.O POSITION 6 138.35 131.16 102.75 106.O0 119.38 119.43 -- 85.-07 ill. 80 114.27 95.91 91.13 81.51 8.6 -EQUILIBRIUMS CONDITIONS ON TEST TRAY PRESSURE ABOVE TRAY, ATM 1.0117 1.0100 1.0030 1.0050 1.02o0 1.0o204 1.01l14 1.0114 1.0241 1.0241 1.0167 1.o201 1.0231 TEMPERATURE, "F 100.1 l O0. O 100.2 100.8 100.4 lO 1 00. 5 1003 lo.4 100.35 o. 1 00.3 1OO.5 100.3 CO PARTIAL PRESSURE, ATM 0.8513 0.8331 0.5661 0.5661 o.6889 o.6890 0.4849 o.4810 0.6117 O. 6130 0.9044 0.8912 0.7299. 02 - 2~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~073 HENRY'S LAW CONSTANT, ATWMOL FR 256.7 256.6 257 258 257.3 257.5 257.2 257,53 257.2 257.3 257.1 257.3 5. x*, MOL FR X 105 331.6 524.7 22-0.3 219.4 2-67.7 267.6 1885.5 186.9 237.8 258.2 351.8 346.428o C*, LB. MOL PER CU. FT. X 105 192.8 188.9 12-8.0 127.7 155.7 155.6 109.6 108.7 138.3 138.1 244 -0. 6. MURPHREE EFFICIENCY % Emi 16. 68.56 66;-25 77.34 79.19 73.97 74.21 77.38 78.23 80.36 38.51 36.90397 Eml 15 57.89 60.05 69,.14 - -68.o02 68.56 68.92 69.95 73. 26 73. 38 31.94 31.09316 ~SAMPLES WITHDRAWN BY USE OF HYPODERMIC SYRINGE, OTHER SAMPLES WITHDRAWN THROUGH VALVE INTO SAMPLE BOTTLE.

T~B~ III-G (CO. minD) PLATE EFFICIENCIES IN RECTANGULA~ COIDMN AT UNIVERSITY OF MICHIGAN, CARBON DIOXIDE-CYCLOHEXANOL SYSTEM wEIR HEIGHt, 2 IN; SPLASH BAFFLE HEIGE~, 2-1/2 IN. RUN 40. 36-A 36-B 30-A 30-B - 43-A 43-B 33-A 33-B 44-A 44-B 50-A 50-B 48-A 48-B BAROMETRIC PRESSURE, IN Po 29.22..... 29422 29.47 29.47 28.88 28.88 29.02 29.02 28.79 28.79 29.33 29.33 29.27 29.27 LIQUID,:~ FLOW ~" ~ ROTAMETER READING' 50.0 50.0 49.5 49.5 50.0 50.0 50.0 50.0 50.0 50.0 50.0 50.0 80.0 80.0 LIQUID TEMP. AT ROTA~TER,~F 100.3 100.1 98.0 97.3 100.7 100.3 101.5 100.3 100.7 100.5 94.9 96.0 100.1 100.1 ~ LB. MOL PER MINUTE 1.288 1.288 1.271 1.270 1.288 1.288 1.288 1.288 1.289 1.289 1.289 1.289 2.062 2.062 GALLONS PER MINUTE 16. 540 16. 550 16.320 16.320 16. 550 16. 540 16. 547 16. 547 16. 550!16.55 16.54 16.54 26.484 26.484 G~LLONS PER MIN. PER FT. WEIR 26.46 26.48 26.11 26. ll 26.48 26.46 26.48 26.48 26.48 26.48 26.46' 26.46 42.37 42.37 GAS FLOW AT ROTAMETER ROTAMETER READING ~ lOO 100 2o5 2o5 24o 24o 29o 29o 3oo 3oo 2o8 2o5 5o 50 TEMPERATURE AT ROTAMETER, ~F 148.0 148.6 131.6 131.6 145.7 -- 158.7 158.7 160.0 160.1 137.6 140,7 130.6 130.9 PRESSURE AT ROTAMETER, IN Hg 32.22 32.22 31.87 31.67 52.-08 32.08 32.92 32.82 32.99 32.99 33.93 53.53 30.87 ~30.87 LB. MOL PER MINUTE 0.2828 0.2820 0.4450 0.4420 0.5040 -- 0.6150 0.6150 O.6190 0.6200 ~0.4547 0.4491 0.2056 0.2062 ~UUBIC FEET PER MINUTE 116.4 116.1 179.8 180.2 207.8 -- 252.4 252.6 253.'0 254.0 174.8 175.6 85.80 86.11 GAS FLOW AT TEST TRAY SUPERFICIAL GAS VELOCITY, [rs, FT/SEC. 2.978 2.970 4.910 4.900 5.590 -- 6.810 6.800 6.780 6.840 4.837 4.816 2.254 2.261 DENSITY, PG LB. PER CU. FT. O. 09957 O. 1002 O. 09562 O. 09495 O. 09395 -- O. 08684 O. 08686 O. 09222 O. 09107 O. lO13 O. 09993 O. lO15 O. 1014 F-FACTOR, Vs~ PG O. 9410 O. 9400 1.497 1. 506 1.710 -- 2. 004 2.004 2.060 2. 064 1.539 1. 522 0.7191 O. 7190 I ~ HYDRAULIC DATA "., O FROTH k~IGHT, IN. 8.0 8.0 -- 11.7 11.7 13.-0 1.3.0 t2.5 12.5 10.5 10.5 9.3 9.3 h) CLEAR LIQUID HEIGPFf, IN POSITION 2 3.5 3.5 3-55 3.70 3.75 3.75 3.80 3.80 5.95 3.95 3.75 3.75 4.60 4.60 i POSi'TION 3 2~55 2.55 2.55 2.65 2.75 2.75 ~2.95 2.95 3.00 ~, 3.00 2.70 2.70 5.60 3.60 i POSITION 4 2'.50 2.50 2.40 12.50 2.55 2-55 2.75 2-75 2.85 2.85 ~2'55 2-55 3.65 3.65 POSITION 5 3.18 3.18 3.40 5.90 3,60 5.60 3'.75 ~ 3.75 3'80 ~ 3.80 3.60 5.60 ~4.25 4.25 r~s~ DROP ACROSS TRA~, IN ~0' -............. 10.0 14.4 14.8 8.40 7.80 4'.40 4.40 GAS COMPOSITION MOL FR y (ENTERING TRAY)...................... 0.6883 -- D.7756 <L~AVI~ ~^~ o.63~0 0.6377 o.637~ o.63~o o.6o78 ~o.6o88 o.~o66 0.4o87 0.5366 o.~8o 0.697o 0.68~7 o.7776 0[~46 LIQUID COMPOSITION, LB. MOL X 105. F~T....... POSITION 1 24.83 24.76 31.63 34.48 25.88 23.85 19.33 23.45 26.05 24.89 26.40 26.27 24.70 -37.96 POSITION 2 30.05 30.06 55.46 49.58 48.02 53.46 -- 46.36 57.89 55.11 47.56 47.87 31.82 33.17 POSITION 3 36.10 36.35 67.16 70.97 65.77 74.51 60.49 48.47 78.77 75.28 65.72 64.47 54.57 35.46 POSITION 4 56.07 56.56 89.47 89.33................ 46.83 49.63 POSITION 5 69.46 68.13 77.21 79.34 79.20 78.37 63.81 60.67 79.62 76.50 72.84 79.23 65.70 65.23 POSITION 6' 80.61 75.95 87.07 87.83 88.76 86.49 66.24 67.66 91.70 84.83 ~ 89.16. 84.50 80.36 79, G2 EQUILIBRIUMS CONDITIONS ON TEST TRAY PRESSURE ABOVE TRAY, ATM 1.0535 1.0535 1.0050 1.0050 0.9987 1.0020 1.0033 1.0033 1.0090 0.9706 1.0421 1.0338 1.Oll7 1.Oll7 TEMPERATCJRE, ~'F lO0.3 99.95 lO0.3 lO0.8 lO0.2 lO0.1 lO0.6 lO0.6 lO0.4 lO0.0 lO0.2 100.2 1DO. 4 lO0.4 CO2 PARTIAL PRESSURE, ATM 0.6690 0.6718 0.6426 0.6363 0.6070 0.6100 ~0.4079 0.4074 0.5414 0.5201 0.7264 0.7047 0.7867 0.7837 HENRY'S LAW CONSTANT, ATM/MOL FR 257.2 256.5 257.1 258.0 257.1 256.7 257.6 257.6 257.3 25.7.3 257.0....257.0 257'3 257.3 x*, MOL FR X 105 260.1 261.9 249.9 246.6 236.1 237.6 158.4 158.2 210.4 202.1 282.6 274.2 ~05.8 304.6 C*, LB. MOL PER CU. FT. X 105 151.2 152.3 145.2 143.4 137.3 138.2 92.O8 92.26 122.4 117.5 164.3 160.0 177.8 177.1.... MURPHREE EFFICIENCY % Eml 16 44.00 40.02 48.82 49.04 56.29 54.65 64.31 64.02 67.95 64.55 45.33 43.56 36.29 34.14 Eml 15 35.22 33.91 40.14 41.26 47.73 47.59 60.94 53.89 55.46 55.58 33.57 39.65 26.69 24.90 ~ SAMPLES WITHDRAWN BY USE OF HYPODERMIC SYRINGE~ OTHER SA~S WITHDRAWN THROUGH VALVE INTO SAMPLE BOTTLE.

~m Tram-s, (CONTnmmD) PLA~ EFFICIENCIES IN RECTANG~ COIiIMN AT UNIVERSITY OF MICHIGAN, CARBON DIOXIDE-CYCI~DHEXANOL SY$~M WEIR ~IG~P, 2 IN; Sl~H BAFFI~ HEIGI~r, 2-1/2 IN. RUN NO. 28-A 28-B 31-A )l-B 46-A 46-B 37-A )7-B 58-A 58-B 59-A 59-B -40-A 40-B BAROMETRIC PRESSURE, IN Ng 29.11 29.11 29.55 29.55 28.94 28.94 29..12 29.12 29.22 29.22 28.92 28.92 29.09 29.09 LIQUID FLOW ROTAMEi~5R READING 80.0 80.O 8040 80.0 15.0 15 ~ 0 15.0 15.0 15.0 15.0 15.0 15.0 15.O 15.0 LIQUID TEMP. AT ROTAMETER,~F 95.9 96.6 100,6 100.8 6~.8 64.7 58,1 58.0 57.8 57.9 58.1 58.1 57.6 57.6 LB. MOL PER MINUTE 2.065 2.062 2.062 2.061 O. 5898 O. 5898 O. 5878 O. 5880 O. 5875 O. 5876 O. 5880 O. 3880 O. )8?7 O. 58?7 GALLONS PER MINUTE 26.410 26.410 26.480 26.48 4.850 4.851 4.890 %.890 4.885 4.887 4.890 4.890 4.887 4.887 GALLONSPER MIN. PER FT. WEIR 42.26 42.26 42.37 42.)7 7.760 7.762 7.824 7.824 7.816 7.819 7.824 7.82'4 7.819 7.819 GAS FLOW AT ROTAME~ER ROTANETER EEADING 107 107 200 200 57.0 57.0 100 lOO 200 200 250 250 )15 )15 TEMPERATURE AT ROTAMETER, ~F 153.0 133.8 132.9 133.6 115.2 115.7 51.) 51.5 61.8 54.1 50.0 50.0 53.1 53.5 PRESSURE AT ROTANETER, IN ~g 51.06 )1.06 )1.70 51.70 51.'14 51'.14 32.02 52.02 52.72 32.67 52.72 32.12 55.19 55.09 LB. MOL PER MINUTE 0.2928 0.2917 0.4420 O.4459 0.2087 O.2101 O.3026 O.5022 0.4750 0.4790 0.5650 O,56OO O.7082 0.7078 CUBIC FEET PER MINUTE 126.2 125.9 180.4 181.4 84.11 84.76 105.4 105.5 165.2 164.5 192.4 194.2 2)8.9 239.6 GAS FLOW AT TEST TRAY SUPERFICIAL GAS VELOCITY~ Vs, FT/SEC. 3.231 3.223 4.880 ~.905 2.094 2.108 2.954 2.949 4.660 4.720 5.700 5.'740 7.215 7.209 DENSITT, $G LB. PER CU. FT. ~ ~ 09173 O. 09224 O. 09198 O. 09127 O. 1168 O. 1158 O. 1109 O. 1112 O. 1046 O. 1040 O. 1032 O. 1034 O. 1002 O. lOO0 F-FACTOR, Vs~PG ~. 9787 O. 9791 1.4801 1.4817 O. 7154 O. 7168 O. 98)4 0.9842 1.508 1.52) 1.8)0 1.84) 2.285 2.280 ~DRAULIC DATA FROTH HEIGHT, IN. 10.7 10.7 12.O 12.0 5.2 5.2 5.7 5.7 7.5 7.5 9.0 9.0 9.2 CLEAR LIQUID BEIGHT, IN POSITION 2 9.2 O 4.85 4.85 5.40 5.40 2.55 2.55 2.)0 2.50 220 2.20 2.20 2.20 2.50 2.)0 I POSITION ) ). 60 ).60 ). 90 ). 90 1.95 1.95 1.80 1.80 1.60 1.60 1.75 1.75 2. O0 2. O0 POSITION % ).50 5.50 ).70 ).70 2.00 2.00 1.85 1.85 1.55 1.55 1.60 1.60 t.80 1.80 POSITION 5 4.25 4.25 4.40 4.40 2.20 2.20 2.20 2.20 2.20 2.20 2.25 2.25 2.55 2.35 PRESSURE DROP ACOROSS TRAY, IN F9.O 5.40 5.)0 7.6 7.5 2.70 2.70 5.10 5.10 5,40 6.00 6.90 7,0 10.4 11.2 GAS COMPOSITION, MOL FR........................ _._ y (LEAVING TRAY) 0.5548 0.550% 0.5%17 0.5241 0.9416 0.9180 0.7254 O.7512 0.5855 0.5858 0.6695 0.67%6 0.5731 0.5696 LIQUID COMPOSITION, I~B. MOL X 105 POSITION 1 27.45 26.~5 29.41 27.95 51.77 50.05 29.15 28.89 28.56 25.12 27.20 50. O8 26.56 29.71 POSITION 2 26.56 51.02 4%.25 %4.42 34.49 33.28 60.48 59.69 6%.98 67.92 67.78 72.07 66.64 67.02 POSITION 5 54.99 34.26 50.18 46.92 72..90 47.66 71.37 70.76 71.18 70.60 87.47 88.92 85.27 -- POSITION % 24.27 47.36.... 93.16 90.4) 81.19 80.73............ POSITION 5 108.75 6).59 6'5.88 68.)5 102.82 76.47 87.49 88.62 89.19 85.28 98.89 102..02 95.87 96.28 POSITION 6 67.26 61.90 71.15 74.75 107.50 -105.97 94.29 96.90 91.85 89.54 105.26 108.26 i01.09 103.40 EQUILIBRl-~ CONDITIONS ON TEST TRAY PRESSURE ABOVE TRAY, ATM 1.0077 1.0080 1.0043 1.00%3 1.0241 1.0241 1.O501 1.0501 1. ~ 1.0%24 1.~0000 1.0000 1.0057 1.0057 TEMPERATURE, ~-F 99- 9 100.4 100.4 100.8 59.32 59.58 58.08 58. O1 57.85 57.92 58.14 58.1% 57.61 57.65 CO2 PARTIAL PRESSURE, ATM 0.5590 0.55%8 0.5441 0.5262 0.9642 O. 93~ O.76.17 0.7678 0.6113 0.6086 0.66.95 0.6746 0.5764 0.5728 HENRY, S LAW CONSTANT. AT~MOL FR 256.4 257.) 257.3 258.0 202.5 202.) 200.5 200.5 200..5 200.5 200.8 200.8 2-00.0 200.0 x*, MOL FR X 105 210.2 215.6 211.5 20%.0 476.6 46~.3 )79.9 582.9 )04.9 30).5 55).5 536.0 288.2 286.4 C*. LB. MOL PER CU. FT. X 105 122.2 125.4 /22.9 118.6 285.7 277.9 227.4 229.2 182.5 181.7 199.8 201.4 172.6 171.5 mm~m~E EFFZC~C~ % Eml 16 41.88 55.91 44.47 5t.49 )0.07 50.92 55.16 54.26 41.47 41.54 45.70 46.15 51.54 52.46 Eml 15 55.91 37.61 )8.88 44.4], 28.25 18.90 29.72 30.07 $9.76 58.78 41.98 42.45 46.55 47.40 ~ SAMPLES WITHDRAWN BY USE OF HYPODERMIC SYRINGE, OT~R SAMPLES WITPIDRAWN THROUGH VALVE iNTO SAMPLE BOTTLE.

TABLE IV-G MAS TRANSFER UNITS AND MASS TRANSFER COEFFICIENTS - ABSORPTION STUDIES Liquid Gas Liquid Weir Liquid Viscosity Velocity Contact Height Rate AL us Time DL x 10~' Run No. Inches gpm lb/ft-hr ft/sec F-Factor - Emv.5 EOG15 N0015 NL15 tL sec kLa sec1 ft2/hr 4A 3-1/S 4.95 60.3 0.89 -0.294 55.64 0.02058 0.0o62 o.o163 0.907 16.19 0.0560 0.471 0.297 4B 4.95 60.0 0.84 0.276 49.55 0.02195 0.0178 0.0180 0.892 16.15 0.0555 0.478 0.255 21A 4.94 59.8 1.75 0.569 105.1 0.01360 0.0112 0.0115 1.188 16.14 0.0756 0.475 0.558 21B 4.94 58.8 1.74 0.568 104.9 0.01609 0.0122 0.0125 1.290.16.17 0.0798 0.484 0.563 5A 4.95 59.5 1.73 0.571 105.1 0.01904 0.0145 0.0146 1.554 16.17 0.0949 0.478 0.454 5B 4.95 60.0 1.70 0.558 100.9 0.01552 0.0129 -0.0150 1.512 16.22 0.0809 0.475 0.571 SA 4.95 58.7 2.68 0.859 176.0 0.01676 0.0114 0.0115 2.024 14.77 0.1570 0.484 0.625 2B 4.95 58.7 2.85 0.901 181.3 0.01844 -0.-0122 0.0125 2.250 14.51 0.1557 0.484 0.699 5A 4.95 59.7 5.57 1.112 199.7 0.01663 0.0108 0.0108 2.157 15.57 0.1615 0.477 0.759 19A 4.95 59.9 5.67 1.207 208.6 0.01165 0.00910 0.00914 1.907 12.85 0.1486. 0.475 0.682 19B 4.95 58.6 3.48 1.141 210.0 0.01185 0.00905 0.00904 1.898 15.20 0.1458 0.486 0.672 8A 3-1/2 4.92 147.1 0.89 0.500 48.7 0.00984 0.00878 0.00883 0.430 16.17 o.o266 0.182 0.197 8B 4.92 145.2 0.84 0.282 45.7 0.01174 0.00982 0.00984 0.450 13.15 0.0279 0.182 0.207 6A 4.92 150.0 1.68 o. 564 91.7 0.00874 0.00720 0.00722 0.662 15.35 0.0425' 0.209 0.292 7A 4.92 131.9 1.67 0.561 91.9 0.00952 0.00816 0.00823 0.756 15.71 0.0481 0.205 0.356 7B 4.92 151.2 1.67 0.561 91.9 0.00942 0.00809 0.00813 o.747 15.71 0.0475 0.207 0.550 9A 4.92 119.1 2.55 0.789 151.1 0.00865 0.00744 0.00742 0.975 15.18 0.0841 0.251 0.422 95 4.92 114.4 2.56 0 790 151.0 0.00894 0.00770 0.00772 1.011 15.18 0.0666 0.252 0.458 16A 4.92 124.6 3.34 1.120 184.0 0.00696 0.00562 0.00561 1.052 15.77 0.0749 0.218 0.508 16B 4.92 123.4 5.54 1.119 184.5 0.00752 0.00645 0.00642 1.184 15.77 0.0860.0.222 0.577 20A 4.91 128.3 3.43 1.142 188.2 0.00698 0.00652 0.00651 1.188 15.54 0.0877 0.212 0.605 20B 4.91 126.8 3.45 1.139 188.7 0.00705 0.00627 0.00651 1.191 15.58 0.0877 0.215 0.599 12A 3-1/2 16.45 125.6 1.71 0.565 28.01 0.01545 0.0117 0.0118 0.350 6.59 0.0516 0.215 0.55 12B 16.46 125.7 1.75 0.562 28.41 0.01286 0.0111 0.0112 0.518 6.38 0.0498 0.221 0.5 10A 16.49 127.0 2.50 0.762 37.70 0.01177 0.0101 0.0102 0.384 6.57 0.0605 0.214 0.412 lOB 16.46 124.2 2.52 0.763 38.06 0.01127 0.00958 0.00944 0.559 6.56 0.0564 0.219 0.581 17A 16.46 150.0 5.59 1.110 55.49 0.00942' o.oo845 0.00843 0.468 6.69 0.0(00 0.209 0.484 17B 16.46 129.5 5.59 1.108 55.58 o.oo998 0.00879. 0.00883 0.'491 6.72 0.X0751 0.210 0.504 D Dffusivities predicted by Wilke-Chang correlation.(13)

305TABIE IV-G (CONT'D) MASS TRANSFER UNIT" AND MASS TRANSFER COEFFICIENTS - ABSORPTION STUDIES * I -T - - - -~, I, T j; *. I z- Liquid Gas Liquid Weir Liquid Viscosity Velocity Contact Height Rate IL Time D x 105 Run No. Inches gpm lb/ft-hr ft/sec F-Factor N kLa sec1 ft/ kDL EMV15 EOP15 OG15 15 le * - - I- I - rei - I r -- r' - - -~~~~~~~~ —— ~I-~~t sec kL sec-1 t2/hr k~/DL1/ 15A 5-1/2 26.9 135.0 0.86 0.287 8.62 0.01626 oo0151 0.0152 0.131 4.84 O.0271 0.201 0.192 15B 27.6 127.8 0.87 0.287 8.51 0.01860 0.0170 0.0171 0.146 4.72 0.0309 0.213 0.212 13A 26.3 107.2 1.74 0.566 18.14 0.01590 - 0.0140 0.0141 0.256 4.67 0.0548 0.258 0.341 13B 26.3 107.0 1.70 - 0.549 17.65 0.01650 0.0144 0.0145 0.256 4.63 0.0553 0.2 8 0.343 11A 26,4 108.9 2.34 0.764 24.34 0.01277 0.0112 0.0113 0.275 4.52 0.0608 0.254 0.382 11B 26.4 106.8 2.33 0.757 24.28 0.02118 O. 0183 0.0185 0.449 4.52 0.0993 0.258 0.617 18A 26.3 125.6 3.40 1.093 34.91 0.0196 O. 0.0112 0.0113 0.94 4.54 0.0868 - 0.215. 0.592 18B 26.3 124.4 3.41 1,095 35.06 0.01131 O.lOl O.010e 0.358 4.54 o.o789 0.218 o.533 26A 2 4.93 59.4 1.08 0.361 65.5 0.00850 0.00692 0.00692 0.453 9.o6 0,0500 0.478 0.229 26B 4.93 59.4 1.09 0.364 65.6 0.00805 o.oo00644 0.02642 0.421 9.00oo o.468 0.478 0.214 23A 4.93 58.9 1.79 0.591 107,7 0.00733 0.00583 0.00581 0.626 8.53 0.0734 0.482 0.334 23B 4.93 58.1 1.78 0.588 107,6 0.00871 0.00710 0,00712 0.766 8.55 o.o896 0.490 0.405 24A 4.94 58.4 2.55 0.847 152.7 0.00700 0.00561 0.00712 0.857 8.05 O.1065 0.487 o,482 24B 4.94 58.4 2.55 0.848 153.1 0.00707 0.00565 0.00561 0.859 8.09 O.1062 0.477 0,481 27A 4.93 57.4 3.09 0.944 185.6 o.00oo3o6 0.00270 0.00270 0.501 7.99 0.0627 0.497 0.281 27B 4.93 56.9 3.17 0.943 186.3 0.00323 0.00287 0,02290 0-.540 7.99 O.0676 0.503 0.301 34A 4.94 58.4 2.99 0.953 180.1 0.00503 0.00409 o.oo4lo 0.738 7.94 0.0929 0.487 0.420 34B 4.94 58.1 2,99 0.953 180.4. 00648 0.00510 0.00511 O.922 7.94 0.1161 0.491 0.524 25A 4.93 59.0 3.10 1.03 186.2 0.00670 0.00526 0.00531 0.989 7.84 0.1261 0o,482 0.574 25B 4.93 58.8 3.11 o.99 187.5' 0.oo694 0.00537 0.00541 1-.014 7.91 0.1282 o0.484 0.584 49A 4.94 58.4 4.29 1.36 259.0 0.00619 0.00453 0.00451 1.168 7.72 0.1513 0.487 0.685 49B 4.94 58.0 4.30 - 136 259.4 0.00568 0.00434 0.00430 1.115 7.72 0.144 0.491 0.653 45A 4.94 59.0 4.43 1.43 267.0 0.00512 0.00385 0.00oo380 1.15 7.76 0.1308 0.482 0,596 45B 4.93 59.2 4.41 1.42 265.4 0.00563 0.00421 0.40420 1.115 7.73 0.1442 0.480 0.658 29A 4.93 58.7 4.90.1.49 294.2 0.00773 0.00538 0.00541 1.592 7.84 0.2031 0.483 0.925 42A 4.94 58.4 5.65 1.77 339.7 0.00577 0.00380 0.00380 1.291 8.49 0.1521 0.488 0.688 32A 4.95 58.6 6,81 2.05 4l10.2 0.00539 0.00377 0.00380 1.559 9.50 0.1641 o,484 o.746 32B 4.94 58.4 6.90 2.07 416.2 O.o00600 o. o0393 0. 00390 1,623 9.57 0.1696 0.487 O.767 41A 4.94 58.6 6.91 2.14 417.0 0.00679 9.88 o.486 * Diffusivities predicted by Wilke-Chang correlation.(l3)

— 306 — TOLE IV-G (CONT'D) MAss TRANSFER UNITS AND MAS TRANSFE13'COEFFIIENTS ABSORPTION STUDIS Liquid Gas Liquid Weir Liquid Viscosity Velocity Contact.,:L~15 ffeight' Rate. L Us Time: D 0 Run'No. Inches gpm lb/ft hr ft/sec F-Factor E ke tFscl f h j/ / EMV15 EOG15 NOG15 1 46A 2 4.84 227.5 2.09 0.715 105'6 0.00342 0.00298 0.00300 0.317 10.48 0.0302 0.116 o.;280 46B 4.85 227.4. 2.11 0.717 1 06.8 O. 0016 O.001l98 O. 00.02 024 9 1. 34 0.116 0.217 37A 4.89 237.3 2.95 0. 983 147.0 mO.0283 O 0 060. 0060 O. 382 9.2 9 O.i 01. 110 O 9 37B 4.89 237.6 2?95 0.984 148.7 O.028002200200.387'.2.041.31.97 38A 4 88 238.0 4.66 1.5i 235.3 O.o002802 O. 0(51 0.002e50 O. 588 8.76 O.o0671 0. ~0964 38B 4.89 237.9 4.72 1.52 237.8 0.o0246 0.00227 9Q.0230' 0.547 8.74 0.o626 O.110 0.599 39A 4.89 236.0 5.0.83 292.4 0.0o246 0.00222 0.00220, 0.643 8,47 0.0759 0.110 0.723 39B 4.89'236+0o 5.74 1, 1.4 289.8 0.00255 0. 00230 0.00230 0.666 8.43 0,0790 O.110 0.752 40A 4.89 238.4 7.21 2.28 363.3 0.0039 0.O214 0.00210 0.763 8.83 oo6.0,3'40B 4.89 238.2 7.21 2.28 363'. 0.002-47.0.00207 0.00210 0.763 8.82 0.0865 018 082 47A i6,53 58,6 2.25 ~~~~~~~0.73 o6 0.1148 0.00933 0.00933 0.371 4.10 0.096' 0.484 O.4lO 47B 16.53 58.4 2.19 0. 714 39.6 0.01l127 0. 00930 0. 00930 0. 371 4.09 O. 0906 O: 490.1 35A 16.54. 58.6 2,46 0.780 44.3 ~0.01034 0.0084 0.084 0.379 4.08 0.o928 0.486 0.420 35B 16.54 58.4 2.47'0.979 44.4 0.01147 0,00925 0.00925 0.414 4.08 O. lO14 0.486 0.459 36A 16.54 58.6 2.98 0.941 55.6 0.01002 0.00794- 0.00794 0.472 3.86 012.8.5 36B 16.55 59.3 2.97 0.940 53.5 0.00950 0.00766 0.00766 O. 4]2 4.02 0. 1025 0.479 0.469 30A 16.32 58.6 4.91 1.5.0 89.6 0. 00739 0. 00639 O. 00639 057 3.98 O. 1448 O o.484 O. 658 sob1 6. 3 57.6 4.90 1.51 89.5 0.00773 -oxoo645 0.065o581 3.8o.14 9 0 450656 50A 16.54 58.8 4.84 1.54 87.1 0.00585 0..00535 0.0055.5 o.465 3.94 0. 1~81 0,483 0.538 50B 16.54 58.8 4.82 1.52 86.7 0.00752 o.oo658 0.0658 0.574 5.5 o.1894 0. o.863 ~3A 1 6.55 58.7 5.59 1.71~ 100,7 0.00898...0.00755 0.00755 0.766 4.05 0.1892 0,4835.6 4SB.16.54 59.1 5.72 1.76 102.9 0.00877 O. 08.08.1.20 19670.8089 33A 16.55 58.1 6.81 2.00 122,8 0.01255 0.01182 0.00118 1.477 4.37 0.338 0,490 1;525 33B 16,55 58,1 6.80 2.00 122.6 0.00951 0.00788 0.00788 0.970 4.37 0,2218 0.490 1.003 44A. 16.55 58.5 6.78 2.06 122.5 0.01010 Q.00983 0. 00893 1.102 4.47 0.2467 0,486 1.16 44B 16.55 58,4 6.84 2.06 123.3 0.0106 0.00885 0.00885 1.101 4.'47 026.8.1 48A 2 26.48 58.4 2.25 0.719 25.4 0.01344 0.01130 0.0113. 0.295 3.46 0.0854 0.486 0.386 48B 26.48. 58. 4 2.26 0. 719 25.4 0.0O12 87 O. 01090 O. 0109 0. 278 3.46 0004O. o486 O, 364 28B 26.41.58.4 3.22 0.979 36.1l 0.01620 0.0.1230 0.0123 0.453 3.46 0.1308 0.486 0.592 31A 26.48 58.4 4.88 1.48o 54.9 0.1660 0.0o1430 0.0143 0.675 3.61 0.1870 0.486 0.846. 3l B 26.48 57.6 4.90.8 55.3 0.o1425 O.Oll6o o.i 016O.651 3.61 O. 1804 0.494 9.811 *Diffusivities predicted by Wilke-Chan g correlation. (133

TABLE T-G WARZEL'S DATA FOR CARBON DIOXIDE ABSORPTION AND DESORPTION USING AIR-WATER SYSTEM WITH TBE GAS PHASE DILUTE IN CARBON DIOXIDE Weir Liquid'C" " C" EO02 Height Rate Us Wsrzel tL by From "C" by NOG L10-2 Run No. Inches gpm ft/sec F-Factor x EMV Correlation E0oG sec Equation (135) Equation (135) From EOG2 L see NL ec' e-116 3-1/2 4.58 1.06 0.29 6.13 4.2 13.63 C-115 2.13 0.57 159.8 2.99 4.9 0O209 12.73 5.12 0.02ll o.2141 3.41 0.268 0.292 C-lll.18 0.86 237.0 1.99 5.6 0.0144 12.66 5 -53 0. 0.014445 3.44 0.271 0.296 0-109 4.72 1.29 351.7 1.32 7.1 0.0102 13.64 5-.97 0.0098 0.009g8 3.45 0.253 0.276 0-117 9.16 1.06 0.29 25.81 6.67 2.9 0.0524 12.36 3.78 0.0550 0.0565 1.46 0.118 0.129 c-114 2,15 0.58 80.62 4.57 3.5 0.0307 12.38 4.92 0.0341 o.o347 2.80 0.226 0.247 C-11O 3.19 0.86 120.3 3,11 3.7 0.0215 7.73 5.34 0.0236 o.3 2.87 0.371 0.405 C-108 4.72 1.29 177.08 2.46 4.9 0.0176 8.o00 5.91 0.0184 0.0186 3.29 0.412 0.44g 0-118 18.31 1.06 0.29 19.88 6.75 2.5 O. 0540 4.79 3.62 0.0574 0.0591 1.18 0.246 0.268! 0-113 2.15 0.58 40.57 5.57 3.0 0.0416 5.22 4.54 0.0%52 o.o463 1.87 0.358 0.391 C) O C-112 3.e8 o.86 59.51 4.95 3.4 O0.0357 5,26 5.13 0.0591 0.0399 2.37 0.451 0.492 C-119 2.0 32.0 1.06 0.29 10.60 8.20 0.0686 a.11 3.55 0.0729 0.0757 0.80 0.195 0.213 C-100 4.58 1.05 0.29 79.23 4.10 4.0 0.0300 9.80 4.19 0.0304 0.0o08 2.44 0.249 0.272 c-98 2.11 0.57 157.2 2.81 4.6 0.0197 8.89 5.05 O. 0202 0.0204 3.21 0.361 O. 394 c-95 5.16 0.87 235.1 1.75 5.3 0.0130 7.79 5.41 0.0130 0.0131 3.08 0.395 0.431 c-92 4.67 1.29 345.0 0.70 6.8 0.0060 7.15 5.35 0.0058 0.0058 2.00 0.279 0.305 C-106 4.62 1.27 348.5 0.88 6.8 0.0073 7.18 5,,56 O.0070 O 0070 2.45 0.341 0. 372 c-99 9.16 1.06 0.29 39.84 4.77 2.3 0.0348 13.63 3.85 0.0388 0.0595 1.58 0.116 0.126 c-97 2.11 O.57 78.8 2.41 2.8 0.0184 4.78 4.40 o.e200 0.0202 m 1.59 0.334 0.364 c-90 5.16 0.87 117.7 2.46.3.3 0.0177 4.77 5.11 0.0195 0.0197 2.32 0.486 0.530 c-89 4.67 1.30 173.8 1.41 4.4 0.0112 4.56 5.36 o.o116 0.0117 2.03 0.445 0.486 0-105 4.62 1.27 174.3 1.50 4.4 0.0118 4.56 5.41 0.0122 0.0125 2.15 0.471 0.514 C-101 18.31 1.05 0.29 19.87 5.31 1.7 0.0415 2.88 3.47 0.VO64 0.0475 0.94 0.327 0.351 c-102 2.11 0.57 39.29 4.1 2.1 0.0o306 308 4.28 0.0350 0.0356 1.40 0.455 0.497 c-94 3.14 0.86 58.56 4.02 2.6 O.0286 3.18 4.93 0.0329 0.034 1.96 0.615 0.671 * DL = 84.09 x 10-6 ft2/hr, by Wilke-Chang correlation.53)

TAB LE M4. (CONIT'D) WARZEL'S. DATA FOR CARBON DIOXIDE ABSORPTION AND DESORPHION USING AIR-WATER SYSTEM WITH THE GAS PHASE DILUTE IN CARBON DIOXIDE Weir Liquid OG 2 R~~N. Height Nate us Warzel tL by From'C" by NOGsc~ kj~a x1 Run No. Inches gpm ft/sec F-Factor x EW Correlation EOG sec Equation (135) Equation (135) From E NL se-2l C-95 4.71 1.31 87.85 2.63 3.5 0.0202 3.29 5.31 0.0218 0.0221 1.94 0.589 0.643 C-107 4.65 1.27 86.31 2.82 3.5 0.0214 3.21 5.35 0.0232 0.0235 2.03 0.631 0.689 C-103 32.0 1.05 0.29 11725 5.92 1.3 0.0478 2.40 3.19 0.0538 0.0553 0.62 0.259 0.283 C-104 2.11 0.57 22.59 5.05 1.7 0.0386 2 93 4.02 0.0445 0.0455 1.03 0.350 0.382 c-96 3.15 0.87 33.54 5.88 2.1 0.0415 3.23 4.78 0.0492 0.0505 1.69 0.524 0.572 CD-85 2.0 4.58 1.04 0.28 78.01 3.99 4.-0 0.0295 8.40 4.16 0.0298 0.0302 2.36 0.281 0.306 -CD-84 3.11 0.84 8 233.26 2.64 5.3 0.0175 7.90 5.74 0.0179 o.0181 4.22.0.535 0.583 CD-122 3.15 0.85 235.4 -2.78 5.3 0.0i81 7.90 5.80 0.0186 0.0188 4.42 0.560 0.611 CD-121.4.68 1.28 346.2 1.97 6.8 0.0136 7.61 6.32 0.0134 0.0134 4.66 0.612 0.668 CD-83 9.16 2.08 0.56 78.27 3.54 2.8 0.0246 4.78 4.67 0.0278 0.0282 2.21 0.462 0.504 CD-82 3.11 0.84 117.3 2.86 3.3 0.0197 4.77 5.22 0.0221 0.0223 2.62 0.549 0.599 CD-120 4.71 1.29 175.9 2.24 4.4 0.016o 5.84 5.81 0.0171 0.0173 3.03 0.519 0.566 CD-86 18.31 1.06 0.28 19.71 4.30 1.7 0.0349' 3.03 3.34 0.0383 0.0391 0.77 0.254 0.277 CD-80 2.09 0..56 39.07 4.05 2.1 0.0302 3.16- 4.25 0.0344 0.0350 1.37 0.434 0.473 CD-79 3.12 0.85 58.47 3.43 2.6 -0.0254 3.26 4.79 0.0287 0.0291 1.70 0.521 0.568 CD-81 4.63 1.27 86.73 2.91 3.5 0.0219 3.29 5.38 0.0239 0.0242 2.09 0.637'0'.695 CD-87 32.0 1.07 0.29 11.47 5.12 1.3 0. 0423 2.40 3.13 0.0469 0.0481 0.55 0.230 0.251 CD-88 3.13 0.85 33.57 4.85 2.1 0.0359 3.23 4.62 0.0416 0.0424 1.42 0.441 0o.480 DL-n,= 84._oq 10 ft /hr, by Wilke-Chang correlation.(13)

-309TABLE VI-G COMPARISON OF POINT EFFICIEECY VALUES DETERMINED BY USE'OF SEVERAL DIFFERENT RELATIONSfIPS BETWEEN POINT AND PLATE EFFICIENCY Weir Liquid Run No., Inches gpm Viscosity, cp ftsea C 7c C n(4) aL_ x x EMV ~00(2 EO() ~5 EG6 Xo - x 3-1/2 4.95 3 37.5 2.68 3.05 2.82 1.01 1.86 1,47 0.0168 0.0117 0.0!24 O.o0972 0.0114 0~CO623 2B 2.83 3.55 3.33 ~99 1.86 1.52 m.0184 0.0130 0.0137 O. l 01.0122, 000810 3A 3.37 3.14 4.65 1.16 1.96 1.55 O.0166 0.0113 0.0131 0.00879 0.0108 0.00734 4A O. 89 2.28 1.95 o.62 2.12 126.o04 o. 0166 o. 0172 O. 0147 O. 0162 O, 0136 4B 0.84 1.79 1.57 0,67 1.99 1,24 0.0220 0.0171 0. 013 0.0154 0. 017U 0.0148 5A 1.73 4.39 4.15 o. 5 1 0.0190 0.0157 -.0145 O. 145 0.0105 5B 1.70 4.52 4.31 0.44 1.50 1.21 0.0155 0.0133 O 0.0123 0.0129 0.00934 19A 3,67 3.47 3.38 o.66 1.50 1.28 O.0116 O,00883 - 0.00676 o.00910 0.00590 19B 3,48 3.24 3.09 0.84 - 1.31 0.0118 0.00879 - 0.00759 o.00905 0.00595 21A 1,75 3.27 3.08 0.55 1.55 1.21 O. 0136 0.0113 0.0117 0.0102 O0 0112 O.00o844 21B 1,74 2.95 2,76 0.69 1.89 1,32 0,0161 0.0127 0.0133 0.0112 0,0122 0.00945 8A 3 1/2 4.92 0.89 1.69!,24 0.52 2.10 1.12 O. 00984 O. 00866 O. 00891 O. 00667 0.00878 O. 0804 8B O.84 1.64 1,20 0.52 - 1.20 0.0117 O. 0102 -O 00982 0. 00940 6A 1.68 - - O.43 o 1.21 o.00874 - - o.00762 0.00720 O.00642 7A 1,67 1.02 - 0.36 - 1,77 o. 00952 m.00687 - o. 0013 O.008 16 o. 00684 7B 1.67 - - O.41 - 1.16 O.00942 - -.00946 0o 00809 o. 00679 9& 2.35 3.20 2.93 o.39 1,56 1.16 O.00865 oo00741 0.00761 0.00707 0,00744 O.m0648 9B 2.36 3.84 3.57 0.39 1.50 1.16 O.00894 0.00780 0,00796 0.00730 0.00770 0.00592 16A 3.34 1.83 1.50 0.56 1.83 1.24 O.00696 0.00527 - 0.00519 0.00562 O. 0448 16B 3.34 4.66 4.45 o.49 1.28 1.13 0.00732 O.x0645 - 0.00568 o.xo64 0O00463 2OA 3.43 4.41 4.17 0.54 1.32 1.10 O.0o698 0.00611 O. 00622 0.00526 O. 00632 O. 00446 20B 3.43 4~10 3.86 O. 52 1.32 1.12 O. 00705 O. 00611 O. 0062 0 o 00537 O. 00627 O. 00450 12A 3,1/2 16.4 ~ 24 1.71 1 o23 0.42 6.40 1,15 0.0134 0.0117 0. 0119 0.0117 o,0114 12B 1.73 1,81 0.53 9.78 1.16 0.0129 0.0117 0.0114 0.0111 0,0110 1OA 2.30 1,40 O.65 4.81 1.17 O.0118 0. 0102 O.0104 O.0101 o. 00974 10B 2.32 O.94 - 373 1.20 0.0113 O.00928 7 O. 00938 O. 00937 17A 3.39 2.18 1.54 - 1.12 O.00942 O.00844 - O.00845 0.00758 17B 3.39 1.93 1.28 2.12 1.14 O.00998 O.00876 O.00893 0.00879 0.00793 l1A 3-1/2 26.4 -25 2.34 1.18 - 2.56 1.14 Oolp8 0.0108 - 0. 0112 0.0108 11B 2.33 - 1.29 2.75 1.16 0.0212 - 0.0180 0.0183 0.0159 13A 1.74 1.12 0.21 130 1. 4 0.0159 0.0142 O.0142 O.ol4o 0.0140 13B 1.70 1.13 0.23 - 1.14 O, 165 O.0147 0o0148 O.0144 0.0145

i~5!V-G {;'[, i':eighl~ Rate ciu i Us) - x'Pir, viscosit-'r, 7 T1 n x`~~~~~~~~~i~ OG(2 E 3OOG5 EGd,.k!j 7Tb. Inches g p iostc?,se C' xp.<!57 ~~~~~~~~~~~~~~~O.$!.0,030- 1.. O. C6. 001? 0.015! 0.17 15B O~~~~~~~~~~ ~ ~ ~ ~ ~~~~~.'- 71.03 -1.09 0,01:: 6 0.1~-0.0170 0,0154 1,~A ~~~~~~~ ~ ~ ~ ~~~~~~~~~3.40 1.47 1.16 0.O!)0 0. Oi00125 -.02 181~ ~~~~~~~~ ~ ~ ~ ~ ~~~ ~ ~~~~.4i i. 49 i.12 0.O'5 C.,0C1i I 0.0!10 i1 C). OiO~, 26A 2 " * 5 7 I C6.i 0. k2 01,52 10.7 1.23 0. 0~55~-,0 0. 00695 O. 00,718 Q. 0064 7 0.00o692 0. 00661 26~ i~~~~~~~~~~~~~~~~~~1 09!.1!i 0.)4 0.0 5.5 1.25 0.00805 CO. 00615 0.00679 0.00o620 O. 0C644 0.00o64;6 25~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~, i,0 ia.0 0.5 365.260,73 0.05 O 0062 0055 O. 005e} 0;00510 23B Z Tt, 2,52 2,16 ~~~~~~~~~~~~~~ ~~~~~. 02 4'$ 5. 5 1.265 0.007571 O. 00752OC03 O.07'O.070.OOi 24A 2 5.0 1.82,3. 212 0.00700 0. 00562 0.00526 0.005!5 0.00561_ 0.oo47!6 ~~~~~~~~~~~~~~~~~~~~~~2 B 2 L 20.2 1,70 0.5F8 2.16 1.25 0.00707 0,00567 O. 005 9L 0.00420 O.00o565 0. 047 27;' 3.0 1.6 0. 34 2,183 t. 15.oo 0.00270 0.00277 0052 0.00270 0. 00524 2 7:", 2. 09 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0.00249 0. 002.o24L27B~~~~~~~~~~~~~~~~~~7 27 2.1.9 O~ 1.6 0112 O. 0032) O. 02,;s4 0, 0025 0. 04 9.08 27B lo ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0.00402 4005 54A 2.99 1.96 i 6'3 0.50 ~~~~~~~~~~~~~~~~~~~~~2.70 1,25 0.00503 0.00414 0.00453 0.005870.4 0 0 005 34A ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0.0050 0.00429 25A ).10 2~~~~~~~~~~~~~~~~~~~~~0 2. 2 - 12O0 2.-_7 i.271 0.00670 0. 007539 O.005,-65 0.00455 0.00526 0 65 25A 5,1- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0.00-4}5 2 5B 5.t2.59 2. 12 0.64 2. 24 ~ 1.29 0.00694 0;'00,520 0.00580 0.00496 O00 p. 00'=4 4 9 A ~~~~~~~~~~~~~~~~~~4.29 2.21! 9) 0.76 2,40 4 357 0. 0061.9 0.00466 oo44 0,0041494"~5 000( 49A O~~~~~~~~ ~~~~~~~~ ~~ ~~~~~~~~~~~~~~~~~~~~.o,6 000415 O. 0O4~i- 55 0.03,R 004347.00 49B 4~~~~~~~~~~~~~~~~~~~ —.20 2 1! 94 0.67 1. 7 i i O.05[ 0~0) O.04i.098. 044.004 45A 4~ a)~~~~~~~~~~~~~~~~~: 2.20 7..96 O. 54 1.72!i. 53 O 0~05~12 O, 00598 O. 00422 O. 005 6 O O 2~ j% 0. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~OOL2 0. 005221,45B b.~l ~~~~~~~~~~~~~ ~~~~~2,~1 2.23 0.70 2.I 3_3, 0060.0045 0. 00 6) O. 005750.02!.0)4 4 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0.00405 ~~~~~~~~~~~~~~~~~~~~4A5.6_5 2.34 2.1Z; 0.725:.2. 1.4452 0(3.C O'~,00o577o; 0.0041,9 0.004k9 29A O.-C, 2.4-( 1~~ ~~~~~~ ~~~~~~~~~~~~~~~~~~~~~. E2 0. o,- 0500' 0.004 C.0 "Y 0.05/.0 0.00519 42A 6.126 5 - 215.3 O 0 ).0~1 01 0.00577 0. 002' 32B 6.90 2.74 2,65.97 2.99 1.5) O, 0600 O,6042720.2049 0.0055[.0095 00)00, 4!A 6. 125. 8101.51.016 0.00279 0.00459- 0,004955 0.005;1000)2 4!B 6. 9 1 2.52 2.50 1.04 2.11 1.0! 0.00657 0.00444 O. 004[~,5 0.00575 0.00516~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.9 274~?65 0.7.c9.045 46A 2 4.5, 4~ 2.0'9 - 0.12 0,28 3. 66 -'u o.i0.014~2 - O.O002 96 0.00297 0.00298 0. 002 92 46 B ~ ~ ~~~ ~~ ~~~~~~~~~~~~ 2,'i. - O.17 0.'27 4.85 1.09 0.002i6 - 0.00197 O. 00188B 0.00198 O.o0194 57A 2.95 2.16 1.68 0.25 1.80 1409 0.00'28) 0.00259 0.00263 0. 002?52 0.0026o 0.025 37B 2.9~5 2.06 i o 58 0,24 1.99 1.10 0. 00288 0. 00'262 Oo&- 0267 O. 0054 O.o00262 O.02, 33A ~~~~~~~~~~~~~~~~~~4,66 2.50 2.0o6 0.33 1.67 1.12 0. 00230 0,00249 0.00254 0. 00~57 0.00251 0. 0021!8 A8 4,72 ).47 3.10 0. 2 1.26 1-08 0.00246 0.00228 0.00230 0. 002 16 0. 00227 0.00193 5 9 A ~~~~~~~~~~~~~~~~~~5.70 2.25 1o90 0.32 1.52 J.-iI 0.00246 0. 002!~ 0.00220 0.00209 0.00222 0.00157( 3 9 B ~~~~~~~~~~~~~~~~~~5.74 2.40 1.98 3.16 1o55 1.11 0.00255 0022 0.22 0021 0025 40^ 7.21 2.47 2.07 O. 40.4 1.12 0. 00259 O. 00205 0. 00210 o. 00194 0021 0017 40A 7.21 2.28 1.85 0.40 2,04 1.19 0.00247 O, 00208 0. 00215 0.00200 0.00207 0.00176

TABLE VI-G (CONT'D) COMPARISON OF POINT EFFICIENCY VALUES'-DETERMINED BY USE OF SEVERAL DIFFERENT RELATIONSHIPS BETWEEN POINT AND PLATE EFFICIENCY Weir Liquid Height Rate Liquid Us * Run No. Inches gpm Viscosity, cp ft/sec C n(4) X () G() G(5) Eo (6) 47A 2 16.5 o358 2.25 1.15 0.37 0.52 - 1.2 0.-0115 o.oo00964.00989 0.00873 0.00933 0.00942 47B 2.'19 1.18 0.41 0.51 165 1.21 0.0113 0.00956 0.00979 0..00865 0.00950 0.00933 35A 2.46 1.25 0.47 0.51 271 1.22 0.010o 0.00879 0.00901 0.00793 0o.oo00848 - 0.00851 35B 2.47 1.25 0.54 0.50 86 1.24 0.0115 o0.00o63 o.o0099o 0.00887 0.00925 0.00926 36A 2.98 1.13 0.31 0.59 9300 1.26 0.0100 0.00820 0.00840 0.00734 0.00794 0.00802 36B 2.97 1.14 0.32 0.56 232 1.24 0.00950 0.00786 0.00804 0.00707 0.00766 0.00768 30A 4.91 2.10 1.52 0.39 2.07 1.16 0.00739 0.00642 0.00739. 00606 0. 00639 0.00566 3OB 4.90 1,50 0.81 0.52 2.68 1.20 0.00773 0.00636 0.00773 0.00589 0.00645 0.00588 50A 4.84 1.84 1.28 0.43 1;74 1.09 0.00585 0.00516 0.00528 0.00468 0.00535 0.00472 50B 4,82 1.74 1.20 0.47 1.96 1.14 0.00752 0.00639 0.00662 0.00588 0.00658 o.00579 43A 5.59 1.71 1.17 0. 56 1.95 1.19 oo00898. 00722 0.00751 0.00670 0.00755 0.o00640 43B 5.72 2.19 1.74 0.49 1.46 1.12 0.00877 0.00735 0.00758 0.0068;2 0,00783. 00626 33A 6.81 - - - 1.10 1.06 0.0126 0.01182 0.00759 33B 6.80 2.60 2.10 0.51 1.79 1.21 0.00951 0.00785 0.00808 0.00730 0.00788 0.o00650 44A 6.78 2.47 2.02 0.54 1.36 1.13 0.0101o 0.00818 0.00847 0.00761 o.oo00893 0.00657 44B 6.84 2.41 1.96 0.57 1.38 1.14 0.0101 0.00811 0.00340 0.00747 0.0085 0.00654 48A 2 26.4 38 2.25 1.21 0.43 0.38. 3030 1.19 0.0134 0.0114 0.0120 0.0111 0.011 0.0116 48B 2.26 1.16 0.32 0.39 - 1.18 0.0129 0.0113 0.0i15 0.0106 O.0109 0.0111 28B 3.22 1.15 0.31 0.47 59000 1.51 0.0162 0.0131 0.0134 0.0127 0.0123 0.0127 51A 4.88 1.68 1.02 0.46 9.41 1.16 0.0166 0.0133 0.0157 0.0131 0.0143 0.0118 31B - 4.90 1.69 1.06 0.59 2.76 1.22 0.0142 0.0117 0.0121 0.0104 0. 0116 O.01i05 (1) EOG =- n ( Er + 1) (2) EOG = + 7 n (1 + Z EMV + 1) (2) OG = -+ 1 (4) EOG = X ()EMV + 1)1/n - 1] EMV (5) OG= xavg - x* (6) EOG =1 On (1EMv + 1)

-312RTABLE VII-G DATA USED IN THE CORRELATION OF THE AVERAGE LIQUID CONCENTRATION ON THE TRAY Weir Liquid Xavg - Height Rate us x x Xavg Run No. Ince PM -"a Run No. Inches gpm L, op ft/sec F-Factor n- XEoG Xo x( )042 (XEOGiO,12 4A 3-1/2 4,95' 37.5 0.89 0.294 1.26 0.251 0.901 1.277 4B 0.84 0.276 1.24 9.215 0,883 1.259 5A 1.73 0. 571 1.31 0.270 1.524 1.245 5B 1.70 0.558 1.21 0. 191 1.302 1.172 21A 1.75 0.569 1.21 0.191 1.477 1.190 21B 1.74 O.568 1.32 0.278 1.281 1,283 2A 2.68 0,859 1.47 0. 86 2.00(6 1.352 2B.283 O. 901 1,52 o.419 2.212 1,382 3A. 3.37 1.112 1.55 0.438 2.157 1.414 19A 3.67 1.207 1.28 0.247 1,898 1,186 19B 3,48 1.141 1.31 0;270 1.900 1.210 8A -1/ 4.9 24 0,89 0o.3o00 1.12 0.113 0.428 1.240 8B 0.84 0.282 1.20 0.182 0.449 1.320 6A 1.68 0. 564 1.21 0.191 0.660 1.271 7A 1.67 O. 561 1.17 0.157 0.750 1.210 7B 1.67 O. 561 1.16. 148 0,744 1.200 9A 2.35 0,789 1.16 Q, 148 0 975 1.162 9B 2.36 0 790 1.16 0.148 1.009 1.159 16A 3.34 1.120 1.24 0.215 1.034 1.234 16B 3.34 1.119 1.13 0.122 1.190 1,107 20A 5.43 1.142 1.10 0. 095 1.196 1.076 20B 5.43 1.139 1.12 0. 115 1.183 1.098 12A 3-1/2 16.4 4 24 - 1.71 O 565 1.1.140 O. 328 1.315 12B 1.. 7T3 0..572 1.16 o. 148 0. 315 1.352 10A 2,39 0.762 1.17 0.157 0.381 1.312 10B 2.32 9.765 1.20 0.182 0.357 1.359 17A 3.39 1. 110 1.12 0.113 0.469 1.228 178 3.39 1.108 1.14 0.131 o.489 1.242 15A 3-1/2 26.4 -25 0.86 0.287 1.08 0.0o7 0.367 1.218 15B 0.87 0.287 1.09 0.086 O. 413 1.220 13A 1.74 0 566 1.14 0.131 0O,253 1.343 13B 1.79. 549 1,14 0.131 0.253 1.343 11A 2.34 0.764 1,14 0.131 0,393 1.275 11B 2.33 0.75'7 1.16 0,148 0.639 1.223

-3 13 - TABLE VII-G (CONT'D) DATA USED IN THE CORRELATION OF THE AVERAGE LIQUID CONICENTRATION ON THE TRAY Weir Lquid Xag - x* Height Rate Us X - x x Run Np. Inches gpm L p, eCP ft/sec F-Factor xi0n *,g XOEOG x.12,~~*CI~~~~r(~~~~~LP~~ C- ~ _ -, _: I, J _. ~ - I 26A 2 4.95 ~- 37 1.08 0.361 1.23 0.207 0.460 1.350 26B 1.09 0.364 1,25 0.223 0. 422 1.387 23A 1.80 0. 591 1.26 0.231 0.628 1.330 23B 1.78. 588 1.23 0.207. 764 1.270 24A 2.55 0.846 1.25 0,223 0.857 1.273 24B 2.55 0.848 1.25 0.223 0.865 1.271 27A 3.09 o.944 1.13 0.122 0.594 1.228 27B 3.17, 0.943 1.12 0.113 0.535 1.208 34A 2.99 0.953 1.23 0.207. 0.737 1.278 34B 2.99 0.953 1.27 0.239 0.920 1.282 25A 3.10 1.034 1.27 0.239 0.979 1.273 25B 3.11 0.990 1,29 0.254 1.007 1.290 49A 4.29 1.362 1.37 O. 315 1.173 1.345 49B 4.30 1.362 1.31 0.270 1.126 1.290 45A 4.43 1.431 1.33 0.286 1.028 1.325 45B 4.41 1.415 1.34 0.292 1.117 1.322 29A 4.90 1.490 1.44 0.365 1.583 1,362 42A 5.63 1.769 1.52 0.419 1.291 1.473 32A 6.81 2.046 1.43 o. 358 1.546 1.357 32B 6.90 2.071 1.53 0.425 1.636 1.441 46A 2 4.88 i..14.5 cp 2.09 O.715 1.15 0.140 9.315 1.320 46B 2.11 0.717 1.09 0.086 0,213 1,310 37A 2.95 0.983 1.09 o.o86 0.385 1.221 37B 2.95 0.984 1.10 0.095 0.390 1,232 38A 4.66 1.508 1.12 0.113 0. 591 1.192 38B 4.72 1.523 1.08 0.077 0.540 1.163 39A 5.70 1 830 1.11 0.104 0.649 1.170 39B 5, 74 1.843 1.11 O.104. 666 1.170 40A 7,21 2.283 1,12 0.113 9.778 1.150 40B 7.21 2.280 1.19 0.174 0.752 1.230

TABLE VII-G ~CoNT' D) DATA USED IN THE CORRELATION OF THE AVERAGE LIQUID CONCENTRATION ON THE TRAY Weir Liquid Xavg - x Height Rate Us -x * Xavg - 3 x x Run No. Inches gpm'L'I cp ft/sec F-Factor xo -n X o EX (EOG)0. 12 47A 2 16.5,-38 2.25 0.733 1.23 0.207 0.473 1.348 47B 2.19 0.7i14 1.21 0.191 0.368 1.365 35A 2.46 0.780 1.22 o.199 0.376 1.372 35B 2.47 0.779 1.24 0.215 0.411 1.380 36A 2.98 o.941 1.26 0.231 - O.426 1.397 56B 2.97 0.940 1.24 0.215 0.410 1-_380 30A 4.91 1.497 1.16 0.148 0.572 1.220 30B 4.90 1.506 1.20 0.182 0.576 1.260 50A 4.84 1.539 1.09 o.086 0o.460 1.194 H50B 4.82 1.522 1.14 0.131 0.570 1.221 43A 5.59 1.710 1.19. 174 O. 760 1.230 43B 5.72 1.756 1.12 0.113 O. 806 1.150 33A 6.81 2.004 1.06 O. 058 1.452 1.108 33B 6.80 2.004 1.21 0.191 0.966 1.213 44A 6.78 2.060o 1.13 0.122 1.094 1.118 44B 6.84 2.064 1.14 0.131 1.091 L 128 48A 2 26.4 _-38 2.25 0.719 1.19 0.174 0.287 1.380 48B 2.26 0.719 i.18 o.166 0.277 1.377 28B 3.22 0.979 1.31 0.270 O.444 1.444 31A 4.88 1.480 1.16 0.148 0.785 1.193 31-B 4.90 1.482 1.22 0,199 -0o642 1.287

BIBLIOGRAPHY 1. American Institute of Chemical Engineers Research Cornmittee, "Third Annual Progress Report" (1955). 2. American Institute of Chemical Engineers Research Committee, "Fourth Annual Progress Report" (1956). 3. American Institute of Chemical Engineers Research Committee, "Final Report on Plate Efficiency" (1958). 4. Anderson J.E., Sc. D. Thesis, Massachusetts Institute of Technology, 1954. 5. ArnoLd, J.H., Ind. Eng. Chem., 22, 1091 (1930). 6. Arnold, J.H., Sc. Do Thesis, Massachusetts Institute of Technology, (1932). 7. Ashby, B.B., Ph.D. Thesis, University of Michigan, 1956. 8. Bromley, LoA., and C*R. Wilke, Ind. Eng. Chem.a 433, 1641-48 (1951). 9. Br onJoAoA, Jr., B.S. Thesis, Massachusetts Institute of Technology, (1954). 10. Calderbank, P.H., Trans Inst... Chem. Engrs., 34, 79 (1956). 11. Carey, J.S., Jo Griswold, W.K. Lewis, and WoH. McAdams, Trans. Am. Inst. Chemn. Engrs., 30 504 (1934). 12. Carlson Tor, J. Am. Chem. Soc., 33, 1027 (1911)0 13, Chang, P. and C.R. Wilke, A.I.Ch.E. Journal, 1, pp. 264-270 (1955). 14. Chilton, T.Ho and A.P. Colburn, d. Eng. Chem., 26, 1183 (1934). 15. Chilton, T.H. and A.P. Colburn, Ind. Eng. Chem., 26, 1183 (1934). 16. Chrisney, J.B., M.S. Thesis, University of California, 1948, 17. Chui J.C., Petroleum Processing, 69 39 (1951)4 18. Chu, J.C., Petroleum Processing, 39 and 154 (1951). 19. Chug J.C., J. Forgrieve, Ro Grosso, S;M. Shah, aznd DF. Othmer, A.I.Ch.E. J1urnal, 3,16 (1957)0 20 Colburn, A.P., et al,, Chem. Eng. Prog. 45, 716 (1949). -315

-31621. Crozier, RoD., PhoD. Thesis, University of Michigan, 19560 220 Danckwerts, Po.,V Ind. Eng. Chem.,.43,.-. 1460 (1951). 23. Danckwerts, PVoV. Chem. Eng. Sci., 2, 1 (1953). 240 Dean, PoCo, Dow Chemical Company, Personal Communication. 250 Dow Chemical Company, "Products of the Dow Chemical Company Their Properties and Uses", Copyright 1955. 26. Drickamer, H.Go and J.R. Bradford, Trans. Am. Inst. Chem. Engrs., 39,9 319 9 (,943). 27. Etherington, L.D., Sc. D. Thesis, Massachusetts Institute of Technology, 1949.e 28. Fairbrother, Fo, S.M. Thesis, Massachusetts Institute of Technology, 1941, 29. Farrell, J.oK and JC, Vyverberg, M.S. Thesis, Massachusetts Institute of Technology, 1940. 30. Foss, AoS., Ph.D. Thesis, University of'Delaware, 1956. 31. Gadwa, J.oA, ScoDo Thesis, Massachusetts Institute of Technology, 1936, 32. Gautreaux, MoF. and H.Eo O'Connell, Chem. Eng. Progo, 51, 232 (1955)o 33. Garner, FoH. and Do Hammerton, Chem. Engo Sci., 3, 1 (1954)o 34. Geddes, RoL., Trans. Am. Inst. Chem. Engrs., 42, 79 (1946). 35. Gerster, J.A., B anet, W.E., and I. Hess, Chemo Eng. Prog., 47, 621 (1951). 36. Ibid 47, 523 (1 951) 37. Ibid, 45, 716 (1949). 38. Gerster, J.A., Memorandum to Members of Technical Subcommittee, AoIoCh.oEo Plate Efficiency Research Program, March 21, 1957. 39. Grohsey E.W., et al,, Chem. Engo. Prog., 45, 725 (1949). 40. Hadamard, J., Cornpt. redo,, 152, 1735 (1911)3 154, 107 (1912). 41. Hatta, S. andd M. Katori, J. Soc. Chemo Indo Japan, 37, 280B (1934). 42. Higbie, R., Tram so AmO In3Lsto Chemo Engrs., 31, 365 (1935)

-31743. Hildebrand, F. B., Methods of Applied Mathematics, Prentice-Hall, Inc., Englewood Cliffs, New York, 1952. 44. Horton, G.F,, S.M. Thesis, Massachusetts Institute of Technology, 1937. 45. Hughes, R.R. and E.R. Gilliland, Chem. Engl Prog., 48, 497 (1952). 46. Hiufner, G., Ann. Physik. Chemo, 6o0_ 134 (1897). 47. International Critical Tables, McGraw Hill Book Co., New York (1931). 48. Johnstone, H.F. and A.D. Singh, Ind Eng. Chem., 29, 286 (1937). 49. Jordan, J., Eo Ackerman, and R.L. Berger, J.A.C.S,i 78, pp. 2979-83 (1956). 50. Kirschbaum, E., Forsch, Gebiete Ing., B5, 245 (1935). 51. Kirschbaum, E., "Distillation and Rectification", p 276, Chemical Pub. Co., New York (1948). 52. Langdon, W.M. and D.B. Keyes, Ind. Eng. Chem,, 35, 464 (1943). 53. Lewis, W.K., Jr., Ind. Eng. Chem., 28, 399 (1936). 54. Lynch, E.J. and C.R. Wilke, A.I.Ch.E. Journal, 1, 9 (1955). 55. Maget, H.J.R., A.I.Ch.Eo Plate Efficiency Research Assistant, University of Michigan (1956). 56. Mayfield, F.D., et al., Ind. Eng. Chem., 44, 2238 (1952). 57. Mehta, J.Jo and R.H. Parekh, M.S. Thesis, Massachusetts Institute of Technology, 1939. 58. Miller, Proc. Roy. Soc. (London), 106 724 (1924). 59. Murphree, E.V., Ind. Eng. Chem., 17, No. 7, Pp. 747-750 (1925). 60. Narsimhan, G. and P.S. Mene, Trans. Indian Inst. Chem. Eng., 5 59 (1952-53). 61. Nord, M., Trans. Am. Inst. Chem. Engrs., 42, 863 (1946). 62. O'Connell, H.E., Transo Am. Inst. Chem. Engrs., 42, 741 (1946). 63. Oholm,. Medd. Nobelinst, 2, 23 (1913). 64. Oliver, E.D. and C.C. Watson, A.I.Ch.E. Journal, 2, No. 1, 18 ( 1956). 65. Peaceman, D oW. and JoE. Vivian, A.I.Ch.E. Joaurnal, 2, No. 4, pp ~ 437-443 (1956).

-31866. Peavy, C.C. and E.Mo Baker, Ind. Eg.o Chem,, 29_ 1056 (1937). 67. Perry, Jo.H. Edo., Chemical Engineers' Handbook, 1st Edition, McGraw-Hill Book Company iac., New York 1934 68, Polich, W.F., Ph.D. Thesis, Carnegie Institute of Technology, 1954. 69. Quayle, OR., Chem. Rev., 53, pp. 439 -589 (1953). 70. Quigleyg C.J.o A.I. Johmson9 and B.Lo Harris, Chem. Eng. Progress Symposium Series, 51,_ No. 16, 31 71, Rhodes, F.Ho and P.G. Slackman, Id. Eng. Chem. 29, 51 (1937)o 72. Robinson,, D,, Ph.D. Thesis,iin progress, University of Delaware. 73. Rush, F.Eo, Jr.,9 and C. Stirba, paper presented at AoIoCh.E. Meeting, Hwnston (May, 1955). 74. Rybczynski, W0, Bull. Acad. Sci., Cracow (A), 40 (1911). 75. Schilling, G.D., G.H. Beyer, and C.C. WatsonL Chem. Eng. Prog., 49, 128 (1953). 76. Sherwood, T.Ko, Absorption and Extraction 1st Edition, MGrawHill Book Company, Inc., New York, 1936. 77~ Sherwood, To.K. and F.AL. Holloway, Trans. Am. Inst. Chem. Engrs., 36, 39 (1940). 78. Sherwood, T.K. and R.L. Pigford,'Absorption and Extraction g 2nd Edition, p. 301, McGraw-Hill Book Company, Inc.,o New York, 1952. 79. Silberman, E., paper presented at Midwest Conference on Fluid and Solid Mechanicss 1956. 80. Simkin, D.J., M.S. Thesis, Uni'versity of California, 1948. 81. Smolin, W-., I. Goldberg, and H. Welsh, Thesis, Polytechnic Institute of Brooklrn-, 1956. 82. Smosky, AoE. and B.F. Dodge, Ind. Engo Chem. 42, 1112 (1950). 83. Spells, KoE. and S. Bakowski, Trans. Inst. Chem. Engrs., 28, 38 (1950). 84. Spells, K.E. and S. Bakowski, Trans. Inst. Chemo Emgrs., 30, 189 (1952)e. 85. Stone, HoLog Sc. D. Thesis, Massachusetts Institulate of Technology, 1953.

-31986. Toor, IHL. and JoM. Marchello, AoIoChoE. Journal, 4, 97 (1958). 87. Uchida, S. and Ko Matsumoto, J. Soc. Chem. Inad (Japan), 39, 224 (1936)o 88. Van Krevelen, D.W. and J.J. Hoftijzer, Chem Eng. Prog., 46, 29 (1950)o 89. Walter, J.F. and T.Kb Sherwood, Indo Eng. Chem., 33, No. 4, pp. 493-501 (1941). 90. Walter, J.F. and ToK. Sherwood, Ind. Engo Chem., 33, 493 (1941). 91. Ward, D., Personal Communication, University of Michigan, 1957. 92, Warzel, L.Ao., Ph.Do Thesis, University of Michigan, 1955. 93. West, F.oB. W.D. Gilbert, and T. Shimizu, Ido. Eng. Chem., 44, No. 10, pp. 2470-78 (1952). 94. West, F.B., W.oD. Gilbert, and T. Shimizu, Ind. Eng. Chem., 44, 2470 (1952). 95. WhartOn, L., B-So Thesis, Massachusetts Institute of Technology 1955 o 96. Whitman, W.G., Chem Met. Eng., 29, No. 4, (1923). 97. Wilke, C.R. and C.Y. Lee, Id. Eng. Chem. 47, 1253 (1955). 98. Winslow, C.E., Jr., Ph.D. Thesis, North Carolina State College, 1955. 99. Sage, B. H., et. al., Ind. Eng. Chem., 48, 275-288, (1956).

UNIVERSITY OF MICHIGAN 31111111015 02223 2212111111 3 901 5 02223 2212